📨 Fix: Handle Oversized Notification Emails
Summary
This PR adds logic to detect and handle oversized notification emails
exceeding Postmark’s 5 MB limit. Instead of retrying indefinitely, the
system now sends a lightweight summary email with key stats and a
dashboard link.
Changes
Added size check in EmailSender.send_templated()
Sends summary email when payload > ~4.5 MB
Prevents infinite retries and queue clogging
Added logs for oversized detection
Fixes#11119
---------
Co-authored-by: Nicholas Tindle <nicholas.tindle@agpt.co>
Co-authored-by: Zamil Majdy <zamil.majdy@agpt.co>
- Resolves#11251
This fixes all the warnings mentioned in #11251, reducing noise and
making our logs and error alerts more useful :)
### Changes 🏗️
- Remove "Block {block_name} has multiple credential inputs" warning
(not actually an issue)
- Rename `json` attribute of `MainCodeExecutionResult` to `json_data`;
retain serialized name through a field alias
- Replace `Path(regex=...)` with `Path(pattern=...)` in
`get_shared_execution` endpoint parameter config
- Change Uvicorn's WebSocket module to new Sans-I/O implementation for
WS server
- Disable Uvicorn's WebSocket module for REST server
- Remove deprecated `enable_cleanup_closed=True` argument in
`CloudStorageHandler` implementation
- Replace Prisma transaction timeout `int` argument with a `timedelta`
value
- Update Sentry SDK to latest version (v2.42.1)
- Broaden filter for cleanup warnings from indirect dependency `litellm`
- Fix handling of `MissingConfigError` in REST server endpoints
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- Check that the warnings are actually gone
- [x] Deploy to dev environment and run a graph; check for any warnings
- Test WebSocket server
- [x] Run an agent in the Builder; make sure real-time execution updates
still work
### Changes 🏗️
- Modifies the LLM block to extract the actual response from the
dictionary returned by the LLM, instead of yielding the entire
dictionary. This addresses
[AUTOGPT-SERVER-6EY](https://sentry.io/organizations/significant-gravitas/issues/6950850822/).
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] After applying the fix, I ran the agent that triggered the Sentry
error and confirmed that it now completes successfully without errors.
---------
Co-authored-by: seer-by-sentry[bot] <157164994+seer-by-sentry[bot]@users.noreply.github.com>
Co-authored-by: Swifty <craigswift13@gmail.com>
Fixes
[AUTOGPT-SERVER-6KZ](https://sentry.io/organizations/significant-gravitas/issues/6976926125/).
The issue was that: Redirect handling strips the URL scheme, causing
subsequent requests to fail validation and hit a 404.
- Ensures the hostname in the URL is properly IDNA-encoded after
validation.
- Reconstructs the netloc with the encoded hostname and preserves the
port if it exists.
This fix was generated by Seer in Sentry, triggered by Craig Swift. 👁️
Run ID: 2204774
Not quite right? [Click here to continue debugging with
Seer.](https://sentry.io/organizations/significant-gravitas/issues/6976926125/?seerDrawer=true)
### Changes 🏗️
**backend/util/request.py:**
- Fixed URL validation to properly preserve port numbers when
reconstructing netloc
- Ensures IDNA-encoded hostname is combined with port (if present)
before URL reconstruction
**Test Results:**
- ✅ Tested request to https://www.target.com/ (original failing URL from
Sentry issue)
- ✅ Status: 200, Content retrieved successfully (339,846 bytes)
- ✅ Port preservation verified for URLs with explicit ports
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Tested request to https://www.target.com/ (original failing URL)
- [x] Verified status code 200 and successful content retrieval
- [x] Verified port preservation in URL validation
<details>
<summary>Example test plan</summary>
- [ ] Create from scratch and execute an agent with at least 3 blocks
- [ ] Import an agent from file upload, and confirm it executes
correctly
- [ ] Upload agent to marketplace
- [ ] Import an agent from marketplace and confirm it executes correctly
- [ ] Edit an agent from monitor, and confirm it executes correctly
</details>
#### For configuration changes:
- [x] `.env.default` is updated or already compatible with my changes
- [x] `docker-compose.yml` is updated or already compatible with my
changes
- [x] I have included a list of my configuration changes in the PR
description (under **Changes**)
<details>
<summary>Examples of configuration changes</summary>
- Changing ports
- Adding new services that need to communicate with each other
- Secrets or environment variable changes
- New or infrastructure changes such as databases
</details>
Co-authored-by: seer-by-sentry[bot] <157164994+seer-by-sentry[bot]@users.noreply.github.com>
Co-authored-by: Swifty <craigswift13@gmail.com>
Removes CLAUDE_3_5_SONNET and CLAUDE_3_5_HAIKU from LlmModel enum, model
metadata, and cost configuration since they are deprecated
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Verify the models are gone from the llm blocks
Categories and Creators where not sanitized in the full text search
### Changes 🏗️
- apply sanitization to categories and creators
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] run tests to check it still works
We're currently seeing errors in the `DatabaseManager` while it's
shutting down, like:
```
WARNING [DatabaseManager] Termination request: SystemExit; 0 executing cleanup.
INFO [DatabaseManager] ⏳ Disconnecting Database...
INFO [PID-1|THREAD-29|DatabaseManager|Prisma-82fb1994-4b87-40c1-8869-fbd97bd33fc8] Releasing connection started...
INFO [PID-1|THREAD-29|DatabaseManager|Prisma-82fb1994-4b87-40c1-8869-fbd97bd33fc8] Releasing connection completed successfully.
INFO [DatabaseManager] Terminated.
ERROR POST /create_or_add_to_user_notification_batch failed: Failed to create or add to notification batch for user {user_id} and type AGENT_RUN: NoneType: None
```
This indicates two issues:
- The service doesn't wait for pending RPC calls to finish before
terminating
- We're using `logger.exception` outside an error handling context,
causing the confusing and not much useful `NoneType: None` to be printed
instead of error info
### Changes 🏗️
- Implement graceful shutdown in `AppService` so in-flight RPC calls can
finish
- Add tests for graceful shutdown
- Prevent `AppService` accepting new requests during shutdown
- Rework `AppService` lifecycle management; add support for async
`lifespan`
- Fix `AppService` endpoint error logging
- Improve logging in `AppProcess` and `AppService`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- Deploy to Dev cluster, then `kubectl rollout restart` the different
services a few times
- [x] -> `DatabaseManager` doesn't break on re-deployment
- [x] -> `Scheduler` doesn't break on re-deployment
- [x] -> `NotificationManager` doesn't break on re-deployment
Updated block costs in `backend/backend/data/block_cost_config.py`:
- **AIShortformVideoCreatorBlock**: Updated from 50 credits to 307
- **AIAdMakerVideoCreatorBlock**: Added cost of 714 credits
- **AIScreenshotToVideoAdBlock**: Added cost of 612 credits
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Verify AIShortformVideoCreatorBlock costs 307 credits when
executed
- [x] Verify AIAdMakerVideoCreatorBlock costs 714 credits when executed
- [x] Verify AIScreenshotToVideoAdBlock costs 612 credits when executed
<!-- Clearly explain the need for these changes: -->
This PR converts Jinja2 TemplateError exceptions to ValueError in the
TextFormatter class to ensure proper error handling and HTTP status code
responses (400 instead of 500).
### Changes 🏗️
<!-- Concisely describe all of the changes made in this pull request:
-->
- Added import for `jinja2.exceptions.TemplateError` in
`backend/util/text.py:6`
- Wrapped template rendering in try-catch block in `format_string`
method (`backend/util/text.py:105-109`)
- Convert `TemplateError` to `ValueError` to ensure proper 400 HTTP
status code for client errors
- Added warning logging for template rendering errors before re-raising
as ValueError
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan: -->
- [x] Verified that invalid Jinja2 templates now raise ValueError
instead of TemplateError
- [x] Confirmed that valid templates continue to work correctly
- [x] Checked that warning logs are generated for template errors
- [x] Validated that the exception chain is preserved with `from e`
#### For configuration changes:
- [x] `.env.default` is updated or already compatible with my changes
- [x] `docker-compose.yml` is updated or already compatible with my
changes
- [x] I have included a list of my configuration changes in the PR
description (under **Changes**)
- Resolves#11226
### Changes 🏗️
- Drop use of `CloudLoggingHandler` which docs state isn't for use in
GKE
- For cloud logging, output only structured log entries to `stdout`
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Test deploy to dev and check logs
Changes to providers blocks to store in db
### Changes 🏗️
- revet change
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
<!-- Put your test plan here: -->
- [x] I have reverted the merge
## Summary
- Fixes database connection warnings in executor logs: "Client is not
connected to the query engine, you must call `connect()` before
attempting to query data"
- Implements resilient database client pattern already used elsewhere in
the codebase
- Adds caching to reduce database load for user context lookups
## Changes
- Updated `get_user_context()` to check `prisma.is_connected()` and fall
back to database manager client
- Added `@cached(maxsize=1000, ttl_seconds=3600)` decorator for
performance optimization
- Updated database manager to expose `get_user_by_id` method
## Test plan
- [x] Verify executor pods no longer show Prisma connection warnings
- [x] Confirm user timezone is still correctly retrieved
- [x] Test fallback behavior when Prisma is disconnected
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Claude <noreply@anthropic.com>
We currently try to re-init the LaunchDarkly client every time a feature flag is checked.
This causes 5 second extra latency on the flag check when LD is down, such as now.
Since flag checks are performed on every block execution, this currently cripples the platform's executors.
- Follow-up to #11221
### Changes 🏗️
- Only try to init LaunchDarkly once
- Improve surrounding log statements in the `feature_flag` module
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- This is a critical hotfix; we'll see its effect once deployed
LaunchDarkly is currently down and it's keeping our executor pods from
spinning up.
### Changes 🏗️
- Wrap `LaunchDarklyIntegration` init in a try/except
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- We'll see if it works once it deploys
## Problem
The YouTube transcription block would fail when attempting to transcribe
videos that only had transcripts available in non-English languages.
Even when usable transcripts existed in other languages, the block would
raise a `NoTranscriptFound` error because it only requested English
transcripts.
**Example video that would fail:**
https://www.youtube.com/watch?v=3AMl5d2NKpQ (only has Hungarian
transcripts)
**Error message:**
```
Could not retrieve a transcript for the video https://www.youtube.com/watch?v=3AMl5d2NKpQ!
No transcripts were found for any of the requested language codes: ('en',)
For this video (3AMl5d2NKpQ) transcripts are available in the following languages:
(GENERATED) - hu ("Hungarian (auto-generated)")
```
## Solution
Implemented intelligent language fallback in the
`TranscribeYoutubeVideoBlock.get_transcript()` method:
1. **First**, tries to fetch English transcript (maintains backward
compatibility)
2. **If English unavailable**, lists all available transcripts and
selects the first one using this priority:
- Manually created transcripts (any language)
- Auto-generated transcripts (any language)
3. **Only fails** if no transcripts exist at all
**Example behavior:**
```python
# Before: Video with only Hungarian transcript
get_transcript("3AMl5d2NKpQ") # ❌ Raises NoTranscriptFound
# After: Video with only Hungarian transcript
get_transcript("3AMl5d2NKpQ") # ✅ Returns Hungarian transcript
```
## Changes
- **Modified** `backend/blocks/youtube.py`: Added try-catch logic to
fallback to any available language when English is not found
- **Added** `test/blocks/test_youtube.py`: Comprehensive test suite
covering URL extraction, language fallback, transcript preferences, and
error handling (7 tests)
- **Updated** `docs/content/platform/blocks/youtube.md`: Documented the
language fallback behavior and transcript priority order
## Testing
- ✅ All 7 new unit tests pass
- ✅ Block integration test passes
- ✅ Full test suite: 621 passed, 0 failed (no regressions)
- ✅ Code formatting and linting pass
## Impact
This fix enables the YouTube transcription block to work with
international content while maintaining full backward compatibility:
- ✅ Videos in any language can now be transcribed
- ✅ English is still preferred when available
- ✅ No breaking changes to existing functionality
- ✅ Graceful degradation to available languages
Fixes#10637
Fixes https://linear.app/autogpt/issue/OPEN-2626
> [!WARNING]
>
> <details>
> <summary>Firewall rules blocked me from connecting to one or more
addresses (expand for details)</summary>
>
> #### I tried to connect to the following addresses, but was blocked by
firewall rules:
>
> - `www.youtube.com`
> - Triggering command:
`/home/REDACTED/.cache/pypoetry/virtualenvs/autogpt-platform-backend-Ajv4iu2i-py3.11/bin/python3`
(dns block)
>
> If you need me to access, download, or install something from one of
these locations, you can either:
>
> - Configure [Actions setup
steps](https://gh.io/copilot/actions-setup-steps) to set up my
environment, which run before the firewall is enabled
> - Add the appropriate URLs or hosts to the custom allowlist in this
repository's [Copilot coding agent
settings](https://github.com/Significant-Gravitas/AutoGPT/settings/copilot/coding_agent)
(admins only)
>
> </details>
<!-- START COPILOT CODING AGENT SUFFIX -->
<details>
<summary>Original prompt</summary>
> Issue Title: if theres only one lanague available for transcribe
youtube return that langage not an error
> Issue Description: `Could not retrieve a transcript for the video
https://www.youtube.com/watch?v=3AMl5d2NKpQ! This is most likely caused
by: No transcripts were found for any of the requested language codes:
('en',) For this video (3AMl5d2NKpQ) transcripts are available in the
following languages: (MANUALLY CREATED) None (GENERATED) - hu
("Hungarian (auto-generated)") (TRANSLATION LANGUAGES) None If you are
sure that the described cause is not responsible for this error and that
a transcript should be retrievable, please create an issue at
https://github.com/jdepoix/youtube-transcript-api/issues. Please add
which version of youtube_transcript_api you are using and provide the
information needed to replicate the error. Also make sure that there are
no open issues which already describe your problem!` you can use this
video to test:
[https://www.youtube.com/watch?v=3AMl5d2NKpQ\`](https://www.youtube.com/watch?v=3AMl5d2NKpQ%60)
> Fixes
https://linear.app/autogpt/issue/OPEN-2626/if-theres-only-one-lanague-available-for-transcribe-youtube-return
>
>
> Comment by User :
> This thread is for an agent session with githubcopilotcodingagent.
>
> Comment by User :
> This thread is for an agent session with githubcopilotcodingagent.
>
> Comment by User :
> This comment thread is synced to a corresponding [GitHub
issue](https://github.com/Significant-Gravitas/AutoGPT/issues/10637).
All replies are displayed in both locations.
>
>
</details>
<!-- START COPILOT CODING AGENT TIPS -->
---
✨ Let Copilot coding agent [set things up for
you](https://github.com/Significant-Gravitas/AutoGPT/issues/new?title=✨+Set+up+Copilot+instructions&body=Configure%20instructions%20for%20this%20repository%20as%20documented%20in%20%5BBest%20practices%20for%20Copilot%20coding%20agent%20in%20your%20repository%5D%28https://gh.io/copilot-coding-agent-tips%29%2E%0A%0A%3COnboard%20this%20repo%3E&assignees=copilot)
— coding agent works faster and does higher quality work when set up for
your repo.
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: ntindle <8845353+ntindle@users.noreply.github.com>
Co-authored-by: Nicholas Tindle <nicholas.tindle@agpt.co>
<!-- Clearly explain the need for these changes: -->
### Need 💡
This PR addresses Linear issue SECRT-1665, which mandates an update to
Linear's OAuth2 implementation. Linear is transitioning from long-lived
access tokens to short-lived access tokens with refresh tokens, with a
deadline of April 1, 2026. This change is crucial to ensure continued
integration with Linear and to support their new token management
system, including a migration path for existing long-lived tokens.
### Changes 🏗️
- **`autogpt_platform/backend/backend/blocks/linear/_oauth.py`**:
- Implemented full support for refresh tokens, including HTTP Basic
Authentication for token refresh requests.
- Added `migrate_old_token()` method to exchange old long-lived access
tokens for new short-lived tokens with refresh tokens using Linear's
`/oauth/migrate_old_token` endpoint.
- Enhanced `get_access_token()` to automatically detect and attempt
migration for old tokens, and to refresh short-lived tokens when they
expire.
- Improved error handling and token expiration management.
- Updated `_request_tokens` to handle both authorization code and
refresh token flows, supporting Linear's recommended authentication
methods.
- **`autogpt_platform/backend/backend/blocks/linear/_config.py`**:
- Updated `TEST_CREDENTIALS_OAUTH` mock data to include realistic
`access_token_expires_at` and `refresh_token` for testing the new token
lifecycle.
- **`LINEAR_OAUTH_IMPLEMENTATION.md`**:
- Added documentation detailing the new Linear OAuth refresh token
implementation, including technical details, migration strategy, and
testing notes.
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Verified OAuth URL generation and parameter encoding.
- [x] Confirmed HTTP Basic Authentication header creation for refresh
requests.
- [x] Tested token expiration logic with a 5-minute buffer.
- [x] Validated migration detection for old vs. new token types.
- [x] Checked code syntax and import compatibility.
#### For configuration changes:
- [ ] `.env.default` is updated or already compatible with my changes
- [ ] `docker-compose.yml` is updated or already compatible with my
changes
- [ ] I have included a list of my configuration changes in the PR
description (under **Changes**)
---
Linear Issue: [SECRT-1665](https://linear.app/autogpt/issue/SECRT-1665)
<a
href="https://cursor.com/background-agent?bcId=bc-95f4c668-f7fa-4057-87e5-622ac81c0783"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://cursor.com/open-in-cursor-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://cursor.com/open-in-cursor-light.svg"><img alt="Open in
Cursor"
src="https://cursor.com/open-in-cursor.svg"></picture></a> <a
href="https://cursor.com/agents?id=bc-95f4c668-f7fa-4057-87e5-622ac81c0783"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://cursor.com/open-in-web-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://cursor.com/open-in-web-light.svg"><img alt="Open in Web"
src="https://cursor.com/open-in-web.svg"></picture></a>
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Nicholas Tindle <ntindle@users.noreply.github.com>
Co-authored-by: Bentlybro <Github@bentlybro.com>
## Summary
Fix critical UserBalance migration and spending issues affecting users
with credits from transaction history but no UserBalance records.
## Root Issues Fixed
### Issue 1: UserBalance Migration Complexity
- **Problem**: Complex data migration with timestamp logic issues and
potential race conditions
- **Solution**: Simplified to idempotent table creation only,
application handles auto-population
### Issue 2: Credit Spending Bug
- **Problem**: Users with $10.0 from transaction history couldn't spend
$0.16
- **Root Cause**: `_add_transaction` and `_enable_transaction` only
checked UserBalance table, returning 0 balance for users without records
- **Solution**: Enhanced both methods with transaction history fallback
logic
### Issue 3: Exception Handling Inconsistency
- **Problem**: Raw SQL unique violations raised different exception
types than Prisma ORM
- **Solution**: Convert raw SQL unique violations to
`UniqueViolationError` at source
## Changes Made
### Migration Cleanup
- **Idempotent operations**: Use `CREATE TABLE IF NOT EXISTS`, `CREATE
INDEX IF NOT EXISTS`
- **Inline foreign key**: Define constraint within `CREATE TABLE`
instead of separate `ALTER TABLE`
- **Removed data migration**: Application creates UserBalance records
on-demand
- **Safe to re-run**: No errors if table/index/constraint already exists
### Credit Logic Fixes
- **Enhanced `_add_transaction`**: Added transaction history fallback in
`user_balance_lock` CTE
- **Enhanced `_enable_transaction`**: Added same fallback logic for
payment fulfillment
- **Exception normalization**: Convert raw SQL unique violations to
`UniqueViolationError`
- **Simplified `onboarding_reward`**: Use standardized
`UniqueViolationError` catching
### SQL Fallback Pattern
```sql
COALESCE(
(SELECT balance FROM UserBalance WHERE userId = ? FOR UPDATE),
-- Fallback: compute from transaction history if UserBalance doesn't exist
(SELECT COALESCE(ct.runningBalance, 0)
FROM CreditTransaction ct
WHERE ct.userId = ? AND ct.isActive = true AND ct.runningBalance IS NOT NULL
ORDER BY ct.createdAt DESC LIMIT 1),
0
) as balance
```
## Impact
### Before
- ❌ Users with transaction history but no UserBalance couldn't spend
credits
- ❌ Migration had complex timestamp logic with potential bugs
- ❌ Raw SQL and Prisma exceptions handled differently
- ❌ Error: "Insufficient balance of $10.0, where this will cost $0.16"
### After
- ✅ Seamless spending for all users regardless of UserBalance record
existence
- ✅ Simple, idempotent migration that's safe to re-run
- ✅ Consistent exception handling across all credit operations
- ✅ Automatic UserBalance record creation during first transaction
- ✅ Backward compatible - existing users unaffected
## Business Value
- **Eliminates user frustration**: Users can spend their credits
immediately
- **Smooth migration path**: From old User.balance to new UserBalance
table
- **Better reliability**: Atomic operations with proper error handling
- **Maintainable code**: Consistent patterns across credit operations
## Test Plan
- [ ] Manual testing with users who have transaction history but no
UserBalance records
- [ ] Verify migration can be run multiple times safely
- [ ] Test spending credits works for all user scenarios
- [ ] Verify payment fulfillment (`_enable_transaction`) works correctly
- [ ] Add comprehensive test coverage for this scenario
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
## Problem
High QPS failures on `spend_credits` operations due to lock contention
from `pg_advisory_xact_lock` causing serialization and seconds of wait
time.
## Solution
Replace PostgreSQL advisory locks with atomic database operations using
CTEs (Common Table Expressions).
### Key Changes
- **Add persistent balance column** to User table for O(1) balance
lookups
- **Atomic CTE-based operations** for all credit transactions using
UPDATE...RETURNING pattern
- **Comprehensive concurrency tests** with 7 test scenarios including
stress testing
- **Remove all advisory lock usage** from the credit system
### Implementation Details
1. **Migration**: Adds balance column with backfill from transaction
history
2. **Atomic Operations**: All credit operations now use single atomic
CTEs that update balance and create transaction in one query
3. **Race Condition Prevention**: WHERE clauses in UPDATE statements
ensure balance never goes negative
4. **BetaUserCredit Compatibility**: Preserved monthly refill logic with
updated `_add_transaction` signature
### Performance Impact
- ✅ Eliminated lock contention bottlenecks
- ✅ O(1) balance lookups instead of O(n) transaction aggregation
- ✅ Atomic operations prevent race conditions without locks
- ✅ Supports high QPS without serialization delays
### Testing
- All existing tests pass
- New concurrency test suite (`credit_concurrency_test.py`) with:
- Concurrent spends from same user
- Insufficient balance handling
- Mixed operations (spends, top-ups, balance checks)
- Race condition prevention
- Integer overflow protection
- Stress testing with 100 concurrent operations
### Breaking Changes
None - all existing APIs maintain compatibility
🤖 Generated with [Claude Code](https://claude.ai/code)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **New Features**
* Enhanced top‑up flows with top‑up types, clearer credit→dollar
formatting, and idempotent onboarding rewards.
* **Bug Fixes**
* Fixed race conditions for concurrent spends/top‑ups, added
integer‑overflow and underflow protection, stronger input validation,
and improved refund/dispute handling.
* **Refactor**
* Persisted per‑user balance with atomic updates for reliable balances;
admin history now prefetches balances.
* **Tests**
* Added extensive concurrency, refund, ceiling/underflow and migration
test suites.
* **Chores**
* Database migration to add persisted user balance; APIKey status
extended (SUSPENDED).
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Swifty <craigswift13@gmail.com>
## Summary
Fixes a critical serialization bug introduced in PR #11187 where
`SafeJson` failed to serialize dictionaries containing Pydantic models,
causing 500 Internal Server Errors in the executor service.
## Problem
The error manifested as:
```
CRITICAL: Operation Approaching Failure Threshold: Service communication: '_call_method_async'
Current attempt: 50/50
Error: HTTPServerError: HTTP 500: Server error '500 Internal Server Error'
for url 'http://autogpt-database-manager.prod-agpt.svc.cluster.local:8005/create_graph_execution'
```
Root cause in `create_graph_execution`
(backend/data/execution.py:656-657):
```python
"credentialInputs": SafeJson(credential_inputs) if credential_inputs else Json({})
```
Where `credential_inputs: Mapping[str, CredentialsMetaInput]` is a dict
containing Pydantic models.
After PR #11187's refactor, `_sanitize_value()` only converted top-level
BaseModel instances to dicts, but didn't handle BaseModel instances
nested inside dicts/lists/tuples. This caused Prisma's JSON serializer
to fail with:
```
TypeError: Type <class 'backend.data.model.CredentialsMetaInput'> not serializable
```
## Solution
Added BaseModel handling to `_sanitize_value()` to recursively convert
Pydantic models to dicts before sanitizing:
```python
elif isinstance(value, BaseModel):
# Convert Pydantic models to dict and recursively sanitize
return _sanitize_value(value.model_dump(exclude_none=True))
```
This ensures all nested Pydantic models are properly serialized
regardless of nesting depth.
## Changes
- **backend/util/json.py**: Added BaseModel check to `_sanitize_value()`
function
- **backend/util/test_json.py**: Added 6 comprehensive tests covering:
- Dict containing Pydantic models
- Deeply nested Pydantic models
- Lists of Pydantic models in dicts
- The exact CredentialsMetaInput scenario
- Complex mixed structures
- Models with control characters
## Testing
✅ All new tests pass
✅ Verified fix resolves the production 500 error
✅ Code formatted with `poetry run format`
## Related
- Fixes issues introduced in PR #11187
- Related to executor service 500 errors in production
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Bentlybro <Github@bentlybro.com>
Co-authored-by: Claude <noreply@anthropic.com>
### Problem
When running multiple backend pods in production, requests can be routed
to different pods causing inconsistent cache states. Additionally, the
current cache implementation in `autogpt_libs` doesn't support shared
caching across processes, leading to data inconsistency and redundant
cache misses.
### Changes 🏗️
- **Moved cache implementation from autogpt_libs to backend**
(`/backend/backend/util/cache.py`)
- Removed `/autogpt_libs/autogpt_libs/utils/cache.py`
- Centralized cache utilities within the backend module
- Updated all import statements across the codebase
- **Implemented Redis-based shared caching**
- Added `shared_cache` parameter to `@cached` decorator for
cross-process caching
- Implemented Redis connection pooling for efficient cache operations
- Added support for cache key pattern matching and bulk deletion
- Added TTL refresh on cache access with `refresh_ttl_on_get` option
- **Enhanced cache functionality**
- Added thundering herd protection with double-checked locking
- Implemented thread-local caching with `@thread_cached` decorator
- Added cache management methods: `cache_clear()`, `cache_info()`,
`cache_delete()`
- Added support for both sync and async functions
- **Updated store caching** (`/backend/server/v2/store/cache.py`)
- Enabled shared caching for all store-related cache functions
- Set appropriate TTL values (5-15 minutes) for different cache types
- Added `clear_all_caches()` function for cache invalidation
- **Added Redis configuration**
- Added Redis connection settings to backend settings
- Configured dedicated connection pool for cache operations
- Set up binary mode for pickle serialization
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
- [x] Verify Redis connection and cache operations work correctly
- [x] Test shared cache across multiple backend instances
- [x] Verify cache invalidation with `clear_all_caches()`
- [x] Run cache tests: `poetry run pytest
backend/backend/util/cache_test.py`
- [x] Test thundering herd protection under concurrent load
- [x] Verify TTL refresh functionality with `refresh_ttl_on_get=True`
- [x] Test thread-local caching for request-scoped data
- [x] Ensure no performance regression vs in-memory cache
#### For configuration changes:
- [x] `.env.default` is updated or already compatible with my changes
- [x] `docker-compose.yml` is updated or already compatible with my
changes (Redis already configured)
- [x] I have included a list of my configuration changes in the PR
description (under **Changes**)
- Redis cache configuration uses existing Redis service settings
(REDIS_HOST, REDIS_PORT, REDIS_PASSWORD)
- No new environment variables required
## Summary
Implement selective rollout of payment functionality using LaunchDarkly
feature flags to enable gradual deployment to pilot users.
- Add `ENABLE_PLATFORM_PAYMENT` flag to control credit system behavior
- Update `get_user_credit_model` to use user-specific flag evaluation
- Replace hardcoded `NEXT_PUBLIC_SHOW_BILLING_PAGE` with LaunchDarkly
flag
- Enable payment UI components only for flagged users
- Maintain backward compatibility with existing beta credit system
- Default to beta monthly credits when flag is disabled
- Fix tests to work with new async credit model function
## Key Changes
### Backend
- **Credit Model Selection**: The `get_user_credit_model()` function now
takes a `user_id` parameter and uses LaunchDarkly to determine which
credit model to return:
- Flag enabled → `UserCredit` (payment system enabled, no monthly
refills)
- Flag disabled → `BetaUserCredit` (current behavior with monthly
refills)
- **Flag Integration**: Added `ENABLE_PLATFORM_PAYMENT` flag and
integrated LaunchDarkly evaluation throughout the credit system
- **API Updates**: All credit-related endpoints now use the
user-specific credit model instead of a global instance
### Frontend
- **Dynamic UI**: Payment-related components (billing page, wallet
refill) now show/hide based on the LaunchDarkly flag
- **Removed Environment Variable**: Replaced
`NEXT_PUBLIC_SHOW_BILLING_PAGE` with runtime flag evaluation
### Testing
- **Test Fixes**: Updated all tests that referenced the removed global
`_user_credit_model` to use proper mocking of the new async function
## Deployment Strategy
This implementation enables a controlled rollout:
1. Deploy with flag disabled (default) - no behavior change for existing
users
2. Enable flag for pilot/beta users via LaunchDarkly dashboard
3. Monitor usage and feedback from pilot users
4. Gradually expand to more users
5. Eventually enable for all users once validated
## Test Plan
- [x] Unit tests pass for credit system components
- [x] Payment UI components show/hide correctly based on flag
- [x] Default behavior (flag disabled) maintains current functionality
- [x] Flag enabled users get payment system without monthly refills
- [x] Admin credit operations work correctly
- [x] Backward compatibility maintained
🤖 Generated with [Claude Code](https://claude.ai/code)
---------
Co-authored-by: Claude <noreply@anthropic.com>
## Summary
Fixes the `Invalid \escape` error occurring in
`/upsert_execution_output` endpoint by completely rewriting the SafeJson
implementation.
## Problem
- Error: `POST /upsert_execution_output failed: Invalid \escape: line 1
column 36404 (char 36403)`
- Caused by data containing literal backslash-u sequences (e.g.,
`\u0000` as text, not actual null characters)
- Previous implementation tried to remove problematic escape sequences
from JSON strings
- This created invalid JSON when it removed `\\u0000` and left invalid
sequences like `\w`
## Solution
Completely rewrote SafeJson to work on Python data structures instead of
JSON strings:
1. **Direct data sanitization**: Recursively walks through dicts, lists,
and tuples to remove control characters directly from strings
2. **No JSON string manipulation**: Avoids all escape sequence parsing
issues
3. **More efficient**: Eliminates the serialize → sanitize → deserialize
cycle
4. **Preserves valid content**: Backslashes, paths, and literal text are
correctly preserved
## Changes
- Removed `POSTGRES_JSON_ESCAPES` regex (no longer needed)
- Added `_sanitize_value()` helper function for recursive sanitization
- Simplified `SafeJson()` to convert Pydantic models and sanitize data
structures
- Added `import json # noqa: F401` for backwards compatibility
## Testing
- ✅ Verified fix resolves the `Invalid \escape` error
- ✅ All existing SafeJson unit tests pass
- ✅ Problematic data with literal escape sequences no longer causes
errors
- ✅ Code formatted with `poetry run format`
## Technical Details
**Before (JSON string approach):**
```python
# Serialize to JSON string
json_string = dumps(data)
# Remove escape sequences from string (BREAKS!)
sanitized = regex.sub("", json_string)
# Parse back (FAILS with Invalid \escape)
return Json(json.loads(sanitized))
```
**After (data structure approach):**
```python
# Convert Pydantic to dict
data = model.model_dump() if isinstance(data, BaseModel) else data
# Recursively sanitize strings in data structure
sanitized = _sanitize_value(data)
# Return as Json (no parsing needed)
return Json(sanitized)
```
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude <noreply@anthropic.com>
### Changes 🏗️
- **Added Claude Haiku 4.5 model support** (`claude-haiku-4-5-20251001`)
- Added model to `LlmModel` enum in
`autogpt_platform/backend/backend/blocks/llm.py`
- Configured model metadata with 200k context window and 64k max output
tokens
- Set pricing to 4 credits per million tokens in
`backend/data/block_cost_config.py`
- **Classic Forge Integration**
- Added `CLAUDE4_5_HAIKU_v1` to Anthropic provider in
`classic/forge/forge/llm/providers/anthropic.py`
- Configured with $1/1M prompt tokens and $5/1M completion tokens
pricing
- Enabled function call API support
### Checklist 📋
#### For code changes:
- [x] I have clearly listed my changes in the PR description
- [x] I have made a test plan
- [x] I have tested my changes according to the test plan:
**Test Plan:**
- [x] Verify Claude Haiku 4.5 model appears in the LLM block model
selection dropdown
- [x] Test basic text generation using Claude Haiku 4.5 in an agent
workflow
#### For configuration changes:
- [x] `.env.default` is updated or already compatible with my changes
- [x] `docker-compose.yml` is updated or already compatible with my
changes
- [x] I have included a list of my configuration changes in the PR
description (under **Changes**)
<details>
<summary>Configuration changes</summary>
- No environment variable changes required
- No docker-compose changes needed
- Model configuration is handled through existing Anthropic API
integration
</details>
https://github.com/user-attachments/assets/bbc42c47-0e7c-4772-852e-55aa91f4d253
---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Bently <Bentlybro@users.noreply.github.com>
## Summary
Move DatabaseError from store-specific exceptions to generic backend
exceptions for proper layer separation, while also fixing store
exception inheritance to ensure proper HTTP status codes.
## Problem
1. **Poor Layer Separation**: DatabaseError was defined in
store-specific exceptions but represents infrastructure concerns that
affect the entire backend
2. **Incorrect HTTP Status Codes**: Store exceptions inherited from
Exception instead of ValueError, causing 500 responses for client errors
3. **Reusability Issues**: Other backend modules couldn't use
DatabaseError for DB operations
4. **Blanket Catch Issues**: Store-specific catches were affecting
generic database operations
## Solution
### Move DatabaseError to Generic Location
- Move from backend.server.v2.store.exceptions to
backend.util.exceptions
- Update all 23 references in backend/server/v2/store/db.py to use new
location
- Remove from StoreError inheritance hierarchy
### Fix Complete Store Exception Hierarchy
- MediaUploadError: Changed from Exception to ValueError inheritance
(client errors → 400)
- StoreError: Changed from Exception to ValueError inheritance (business
logic errors → 400)
- Store NotFound exceptions: Changed to inherit from NotFoundError (→
404)
- DatabaseError: Now properly inherits from Exception (infrastructure
errors → 500)
## Benefits
### ✅ Proper Layer Separation
- Database errors are infrastructure concerns, not store-specific
business logic
- Store exceptions focus on business validation and client errors
- Clean separation between infrastructure and business logic layers
### ✅ Correct HTTP Status Codes
- DatabaseError: 500 (server infrastructure errors)
- Store NotFound errors: 404 (via existing NotFoundError handler)
- Store validation errors: 400 (via existing ValueError handler)
- Media upload errors: 400 (client validation errors)
### ✅ Architectural Improvements
- DatabaseError now reusable across entire backend
- Eliminates blanket catch issues affecting generic DB operations
- All store exceptions use global exception handlers properly
- Future store exceptions automatically get proper status codes
## Files Changed
- **backend/util/exceptions.py**: Add DatabaseError class
- **backend/server/v2/store/exceptions.py**: Remove DatabaseError, fix
inheritance hierarchy
- **backend/server/v2/store/db.py**: Update all DatabaseError references
to new location
## Result
- ✅ **No more stack trace spam**: Expected business logic errors handled
properly
- ✅ **Proper HTTP semantics**: 500 for infrastructure, 400/404 for
client errors
- ✅ **Better architecture**: Clean layer separation and reusable
components
- ✅ **Fixes original issue**: AgentNotFoundError now returns 404 instead
of 500
This addresses the logging issue mentioned by @zamilmajdy while also
implementing the architectural improvements suggested by @Pwuts.
## Summary
Fix store exception hierarchy to prevent ERROR level stack trace spam
for expected business logic errors and ensure proper HTTP status codes.
## Problem
The original error from production logs showed AgentNotFoundError for
non-existent agents like autogpt/domain-drop-catcher was:
- Returning 500 status codes instead of 404
- Generating ERROR level stack traces in logs for expected not found
scenarios
- Bypassing global exception handlers due to improper inheritance
## Root Cause
Store exceptions inherited from Exception instead of ValueError, causing
them to bypass the global ValueError handler (400) and fall through to
the generic Exception handler (500) with full stack traces.
## Solution
Create proper exception hierarchy for ALL store-related errors by
making:
- MediaUploadError inherit from ValueError instead of Exception
- StoreError inherit from ValueError instead of Exception
- Store NotFound exceptions inherit from NotFoundError (which inherits
from ValueError)
## Changes Made
1. MediaUploadError: Changed from Exception to ValueError inheritance
2. StoreError: Changed from Exception to ValueError inheritance
3. Store NotFound exceptions: Changed to inherit from NotFoundError
## Benefits
- Correct HTTP status codes: Not found errors return 404, validation
errors return 400
- No more 500 stack trace spam for expected business logic errors
- Clean consistent error handling using existing global handlers
- Future-proof: Any new store exceptions automatically get proper status
codes
## Result
- AgentNotFoundError for autogpt/domain-drop-catcher now returns 404
instead of 500
- InvalidFileTypeError, VirusDetectedError, etc. now return 400 instead
of 500
- No ERROR level stack traces for expected client errors
- Proper HTTP semantics throughout the store API
## Summary
Fix critical SafeJson function to properly sanitize JSON-encoded Unicode
escape sequences that were causing PostgreSQL 22P05 errors when null
characters in web scraped content were stored as "\u0000" in the
database.
## Root Cause Analysis
The existing SafeJson function in backend/util/json.py:
1. Only removed raw control characters (\x00-\x08, etc.) using
POSTGRES_CONTROL_CHARS regex
2. Failed to handle JSON-encoded Unicode escape sequences (\u0000,
\u0001, etc.)
3. When web scraping returned content with null bytes, these were
JSON-encoded as "\u0000" strings
4. PostgreSQL rejected these Unicode escape sequences, causing 22P05
errors
## Changes Made
### Enhanced SafeJson Function (backend/util/json.py:147-153)
- **Add POSTGRES_JSON_ESCAPES regex**: Comprehensive pattern targeting
all PostgreSQL-incompatible Unicode and single-char JSON escape
sequences
- **Unicode escapes**: \u0000-\u0008, \u000B-\u000C, \u000E-\u001F,
\u007F (preserves \u0009=tab, \u000A=newline, \u000D=carriage return)
- **Single-char escapes**: \b (backspace), \f (form feed) with negative
lookbehind/lookahead to preserve file paths like "C:\\file.txt"
- **Two-pass sanitization**: Remove JSON escape sequences first, then
raw characters as fallback
### Comprehensive Test Coverage (backend/util/test_json.py:219-414)
Added 11 new test methods covering:
- **Control character sanitization**: Verify dangerous characters (\x00,
\x07, \x0C, etc.) are removed while preserving safe whitespace (\t, \n,
\r)
- **Web scraping content**: Simulate SearchTheWebBlock scenarios with
null bytes and control characters
- **Code preservation**: Ensure legitimate file paths, JSON strings,
regex patterns, and programming code are preserved
- **Unicode escape handling**: Test both \u0000-style and \b/\f-style
escape sequences
- **Edge case protection**: Prevent over-matching of legitimate
sequences like "mybfile.txt" or "\\u0040"
- **Mixed content scenarios**: Verify only problematic sequences are
removed while preserving legitimate content
## Validation Results
- ✅ All 24 SafeJson tests pass including 11 new comprehensive
sanitization tests
- ✅ Control characters properly removed: \x00, \x01, \x08, \x0C, \x1F,
\x7F
- ✅ Safe characters preserved: \t (tab), \n (newline), \r (carriage
return)
- ✅ File paths preserved: "C:\\Users\\file.txt", "\\\\server\\share"
- ✅ Programming code preserved: regex patterns, JSON strings, SQL
escapes
- ✅ Unicode escapes sanitized: \u0000 → removed, \u0048 ("H") →
preserved if valid
- ✅ No false positives: Legitimate sequences not accidentally removed
- ✅ poetry run format succeeds without errors
## Impact
- **Prevents PostgreSQL 22P05 errors**: No more null character database
rejections from web scraping
- **Maintains data integrity**: Legitimate content preserved while
dangerous characters removed
- **Comprehensive protection**: Handles both raw bytes and JSON-encoded
escape sequences
- **Web scraping reliability**: SearchTheWebBlock and similar blocks now
store content safely
- **Backward compatibility**: Existing SafeJson behavior unchanged for
legitimate content
## Test Plan
- [x] All existing SafeJson tests pass (24/24)
- [x] New comprehensive sanitization tests pass (11/11)
- [x] Control character removal verified
- [x] Legitimate content preservation verified
- [x] Web scraping scenarios tested
- [x] Code formatting and type checking passes
🤖 Generated with [Claude Code](https://claude.ai/code)
---------
Co-authored-by: Claude <noreply@anthropic.com>
