Skip to content

Releases: VirtualFlyBrain/owl_cache

1.1.12

22 Dec 12:18

Choose a tag to compare

1.1.11

15 Dec 17:29

Choose a tag to compare

Full Changelog: 1.1.10...1.1.11

1.1.10

09 Dec 09:56
6328ed4

Choose a tag to compare

Full Changelog: 1.1.9...1.1.10

1.1.9

09 Dec 09:49
1456d9f

Choose a tag to compare

Full Changelog: 1.1.8...1.1.9

1.1.8

08 Dec 22:41
4ce7988

Choose a tag to compare

Full Changelog: 1.1.7...1.1.8

1.1.7

08 Dec 22:03
588ae98

Choose a tag to compare

Full Changelog: 1.1.5...1.1.7

1.1.6

08 Dec 21:54
66daaa8

Choose a tag to compare

Full Changelog: 1.1.5...1.1.6

1.1.5

08 Dec 14:10
f6e5497

Choose a tag to compare

Full Changelog: 1.1.4...1.1.5

Release 1.1.4

12 Nov 17:45

Choose a tag to compare

Release Notes - Version 1.1.4

Overview

This release significantly improves the NGINX caching proxy's ability to handle complex OWL reasoning queries that can take several minutes to complete. Key improvements include extended timeout settings, fixed health endpoint behaviour, and enhanced cache reliability.

🚀 New Features

Extended Timeout Support

  • 900-second timeouts (15 minutes) for proxy connections, reads, and sends
  • Supports complex OWL reasoning queries that can take several minutes to complete
  • Tested with adult brain queries ([FBbt_00003624]) that complete in ~54 seconds

Improved Health Endpoint

  • Direct health responses: Health endpoint now returns immediate "OK" instead of proxying to upstream
  • Faster health checks: Sub-millisecond response times for health monitoring
  • Reliable monitoring: Health checks no longer depend on upstream server availability

🐛 Bug Fixes

Cache Directory Creation

  • Fixed cache initialisation: Dockerfile now properly creates /var/cache/nginx/owlery directory
  • Permissions handling: Correct ownership and permissions for the cache directory
  • Eliminates cache failures: Prevents "cache directory not found" errors on container startup

Health Endpoint Reliability

  • Removed proxy dependency: Health endpoint no longer fails when upstream is unavailable
  • Consistent responses: Always returns HTTP 200 with "OK" for healthy proxy instances
  • Monitoring compatibility: Works correctly with Docker health checks and orchestration systems

🔧 Technical Improvements

Timeout Configuration

proxy_connect_timeout 900s;
proxy_read_timeout 900s;
proxy_send_timeout 900s;
  • Increased from 90 seconds to 900 seconds across all proxy operations
  • Handles long-running OWL reasoning queries without premature timeouts
  • Maintains connection stability for complex anatomical queries

Full Changelog: 1.1.3...1.1.4

Release 1.1.3: Reliable Environment Variable Handling & CI/CD Stability

12 Nov 16:16

Choose a tag to compare

What's New
🛠️ Improvements
Reliable Environment Variable Injection:

Updated Docker and GitHub Actions workflow to ensure all required environment variables (UPSTREAM_SERVER, CACHE_MAX_SIZE, DNS_RESOLVER) are always available inside the container at runtime.
Uses both env: and -e flags in CI to guarantee envsubst generates a valid nginx.conf every time.
CI/CD Stability:

Fixed intermittent CI failures caused by missing or improperly injected environment variables.
Now the container always starts with a valid NGINX configuration in both local and GitHub Actions environments.
🐛 Bug Fixes
Prevents "no 'events' section in configuration" errors by ensuring the NGINX config is always generated with all required variables.
Simplified DNS resolver to a single value (8.8.8.8) for maximum compatibility in CI.
📚 Documentation
README and workflow comments updated to clarify environment variable handling and troubleshooting steps.