1.7.7

What's Changed

• Update cryptography requirement from <44.0.0 to <45.0.0 by @dependabot in #380
• Feature/E-2131 Utility function for resolving model-cache paths from Huggingface repositories by @gabewillen in #377
• fix: streamed errors were previously swallowed by @deanq in #384

Full Changelog: 1.7.6...1.7.7

1.7.6

What's Changed

• Fix: JobScaler issues that cause request failures by @deanq in #383
• Fixes issue #373 for required input validation by @gabewillen in #379

New Contributors

• @gabewillen made their first contribution in #379

Full Changelog: 1.7.5...1.7.6

1.7.5

Recommend Upgrade to 1.7.6

What's Changed

• Fix: failed requests due to race conditions in the job queue vs job progress by @deanq in #376

Full Changelog: 1.7.4...1.7.5

1.7.4

Recommend Upgrade to 1.7.6

What's Changed

• fix: long-running jobs crash with SIGTERM by @deanq in #370
• Bump pylint from 3.2.5 to 3.3.1 by @dependabot in #361
• minimum network speed parameters for create_pod by @dxqbYD in #360

New Contributors

• @dxqbYD made their first contribution in #360

Full Changelog: 1.7.3...1.7.4

1.7.3

Recommend Upgrade to 1.7.6

SDK 1.7.3 Advisory: Known Issues with Long-Running Jobs

1.7.3: Long-running jobs (>60 seconds) can cause the system to stop the worker, triggering retries and failures. Additionally, a long idle timeout (20+ seconds) may result in similar behavior, especially for the second request.

What's Changed

• Refactored rp_job.get_job to work well under pause and unpause conditions. More debug lines too.
• Refactored rp_scale.JobScaler to handle shutdowns where it cleans up hanging tasks and connections gracefully. Better debug lines.
• Fixed rp_scale.JobScaler from unnecessary long asyncio.sleeps made before considering the blocking get_job calls.
• Improved worker_state's JobProgress and JobsQueue to timestamp when jobs are added or removed.
• Incorporated the lines of code in worker.run_worker into rp_scale.JobScaler where it belongs and simplified to job_scaler.start()
• Fixed non-error logged as errors in tracer
• Updated unit tests mandating these changes* Blocking job take call means 5-sec debounce no longer needed by @deanq in #366
• Debounce at HTTP 429 response by @deanq in #367

Full Changelog: 1.7.2...1.7.3

1.7.2

Recommend Upgrade to 1.7.6

SDK 1.7.2 Advisory: Delay Issue on First Message

1.7.2: A delay issue affects the first message sent to the endpoint, with users experiencing lower traffic being impacted more.

What's Changed

• Corrected job_take_url by @deanq in #359
• Update cryptography requirement from <43.0.0 to <44.0.0 by @dependabot in #353
• fix: pings were missing requestIds since the last big refactor by @deanq in #362

Full Changelog: 1.7.1...1.7.2

1.7.1

Recommend Upgrade to 1.7.6

SDK 1.7.1 Advisory: Health Check Failures on Long Jobs

1.7.1: Jobs running longer than 60 seconds may cause the health check to fail, leading to retries and job timeouts.

What's Changed

• Feat: use 'black' to unify formatting by @ef0xa in #348
• Fix: get rid of pylint by @ef0xa in #351
• Make core envvars backwards compatible by @galen-runpod in #354
• Use less strict numbers for coverage. by @galen-runpod in #357
• Add wrappers for UpdateRegistryAuth and UpdateRegistryAuth. by @pconstr in #355
• Batch Concurrency by @deanq in #358

New Contributors

• @ef0xa made their first contribution in #346
• @galen-runpod made their first contribution in #354
• @pconstr made their first contribution in #355

Full Changelog: 1.7.0...1.7.1

1.7.0

Recommend Upgrade to 1.7.6

What's Changed

• Allow newer aiohttp versions by @fastjames in #323
• Observability: implement tracing for http requests by @deanq in #329
• Deprecated JobScaler.concurrency_modifier by @deanq in #331

Full Changelog: 1.6.2...1.7.0

1.6.2

Recommend Upgrade to 1.7.6

SDK 1.6.2 Advisory: Network issues under high rate of requests

1.6.2: Endpoints with high throughputs or concurrency modifiers can experience HTTP/502 from Cloudflare causing bottlenecks and delays on job queues.

Fixed

• Reorder FastAPI endpoints for clarity.
• Truncate long logs to avoid overloading the server.

What's Changed

• Fastai clarity by @justinmerrell in #298

Full Changelog: 1.6.1...1.6.2

1.6.1

Added

• User-Agent for better analytics tracking.
• Expose the ability to create container auth.
• Emulate webhooks for local testing when using the rp_serve_api flag.

Changed

• Updated aiohttp from 3.9.2 to 3.9.3
• [BREAKING] rename registry_auth to registry_auth_id for clarity.
• Added additional details to the FastAPI page.

What's Changed

• User agent tracking by @justinmerrell in #290
• build(deps): bump aiohttp[speedups] from 3.9.2 to 3.9.3 by @dependabot in #292
• feat: added runpod command to create container registry auth and rena… by @yamanahlawat in #294
• 280 enable webhooks for local testing of serverless workers by @justinmerrell in #295
• 1.6.1 cleanup by @justinmerrell in #296

New Contributors

• @yamanahlawat made their first contribution in #294

Full Changelog: 1.6.0...1.6.1