--- title: "#reliability articles | Hashnode" description: "Explore 270 articles tagged #reliability on Hashnode. Tutorials, guides, and how-tos from the community." url: "https://hashnode.com/tag/reliability" --- Tag # #reliability 270 posts · 7 followers [Yilia Lin](https://hashnode.com/@Yilia) · [1d ago](https://hashnode.com/posts/partial-api-outages-need-retry-budgets-not-retry-storms/6af2f9e9e8553480aa21edcc) ### [Partial API Outages Need Retry Budgets, Not Retry Storms](https://api7.hashnode.dev/partial-api-outages-need-retry-budgets-not-retry-storms?utm_source=hashnode&utm_medium=feed) Originally published by Yilia Lin on API7.ai on November 10, 2026. This syndication copy was prepared with AI assistance. Original drafting provenance has not been verified. During a partial outage, s 9 min read · [api7.hashnode.dev](https://api7.hashnode.dev/partial-api-outages-need-retry-budgets-not-retry-storms?utm_source=hashnode&utm_medium=feed) [LastMileOps](https://hashnode.com/@lastmileops) · [2d ago](https://hashnode.com/posts/choose-the-recovery-task-before-you-test-the-backup/6af1a5e7cfbcd6e7d405bc67) ### [Choose the recovery task before you test the backup](https://lastmileops.hashnode.dev/choose-the-recovery-task-before-you-test-the-backup?utm_source=hashnode&utm_medium=feed) Imagine a small internal reporting service. Its backups finish every night, the repository is readable, and yesterday's snapshot can restore a CSV. Then the original machine disappears. The replacemen 4 min read · [lastmileops.hashnode.dev](https://lastmileops.hashnode.dev/choose-the-recovery-task-before-you-test-the-backup?utm_source=hashnode&utm_medium=feed) [Powerduck](https://hashnode.com/@powerduck) · [6d ago](https://hashnode.com/posts/agent-backoff-no-jitter-ignored-retry-after/6aecaf7854116af4fd8e28e4) ### [The Agent Added Backoff but Ignored Retry-After, and Every Client Retried Together](https://blogs.powerduck.com/agent-backoff-no-jitter-ignored-retry-after?utm_source=hashnode&utm_medium=feed) Exponential backoff without jitter makes every client wait the same amount and retry in lockstep, turning one throttled 429 into a synchronized retry storm. The agent also ignored the Retry-After header and baked a generic 5xx retry into a throttle response it never inspected. 4 min read · [blogs.powerduck.com](https://blogs.powerduck.com/agent-backoff-no-jitter-ignored-retry-after?utm_source=hashnode&utm_medium=feed) [Revana](https://hashnode.com/@revana) · [6d ago](https://hashnode.com/posts/video-generation-pipeline-fallback/6aec63fa8fe3a91e6d6aacfc) ### [Building a video-gen pipeline that doesn't fall over: fallback across providers](https://revana.hashnode.dev/video-generation-pipeline-fallback?utm_source=hashnode&utm_medium=feed) Routing, retries, idempotency keys, and only paying for generations that succeed. 9 min read · [revana.hashnode.dev](https://revana.hashnode.dev/video-generation-pipeline-fallback?utm_source=hashnode&utm_medium=feed) [Powerduck](https://hashnode.com/@powerduck) · [Nov 1](https://hashnode.com/posts/retry-post-after-timeout-duplicated-record/6ae785df29f309c9602b16c5) ### [The Retry After a Timeout Created Two Records Because the First POST Had Already Landed](https://blogs.powerduck.com/retry-post-after-timeout-duplicated-record?utm_source=hashnode&utm_medium=feed) A timeout does not prove the server failed. The agent added an automatic retry to a POST on timeout, and the first request had already committed, so the second created a duplicate. Retry logic belongs on idempotent calls, or it needs an idempotency key. 4 min read · [blogs.powerduck.com](https://blogs.powerduck.com/retry-post-after-timeout-duplicated-record?utm_source=hashnode&utm_medium=feed) [Matthew](https://hashnode.com/@matthewmodelfla) · [Nov 1](https://hashnode.com/posts/ai-api-idempotency/6ae69ebf904185065c8b981a) ### [Make AI API Operations Idempotent](https://matthewmodelflare.hashnode.dev/ai-api-idempotency?utm_source=hashnode&utm_medium=feed) Logical request IDs, idempotency keys, request digests, and durable results. 6 min read · [matthewmodelflare.hashnode.dev](https://matthewmodelflare.hashnode.dev/ai-api-idempotency?utm_source=hashnode&utm_medium=feed) [delay.run](https://hashnode.com/@delayrun) · [Oct 31](https://hashnode.com/posts/dead-letter-queues-for-http-jobs-what-they-are-and-how-to-use-one/6ae672fa2ffc6226c1a74302) ### [Dead letter queues for HTTP jobs: what they are and how to use one](https://delayrun.hashnode.dev/dead-letter-queues-for-http-jobs-what-they-are-and-how-to-use-one?utm_source=hashnode&utm_medium=feed) Retries fix most failures. An endpoint was restarting, a database was briefly locked, a third-party API hit a rate limit: try again a few minutes later and it works. But some jobs fail every time, no 4 min read · [delayrun.hashnode.dev](https://delayrun.hashnode.dev/dead-letter-queues-for-http-jobs-what-they-are-and-how-to-use-one?utm_source=hashnode&utm_medium=feed) [Rabin Adhikari](https://hashnode.com/@softverse) · [Oct 31](https://hashnode.com/posts/kubernetes-pod-disruption-budgets-in-2026-graceful-shutdown-without-dropped-requests/6ae5c92baed618e81d91e6d6) ### [Kubernetes Pod Disruption Budgets in 2026: Graceful Shutdown Without Dropped Requests](https://devopsenginer.hashnode.dev/kubernetes-pod-disruption-budgets-in-2026-graceful-shutdown-without-dropped-requests?utm_source=hashnode&utm_medium=feed) Kubernetes Pod Disruption Budgets in 2026: Graceful Shutdown Without Dropped Requests TL;DR: Kubernetes constantly disrupts your pods on purpose — node drains, cluster upgrades, autoscaler scale-down 11 min read · [devopsenginer.hashnode.dev](https://devopsenginer.hashnode.dev/kubernetes-pod-disruption-budgets-in-2026-graceful-shutdown-without-dropped-requests?utm_source=hashnode&utm_medium=feed) [Matthew](https://hashnode.com/@matthewmodelfla) · [Oct 30](https://hashnode.com/posts/ai-api-backpressure/6ae3fd395597d31be248b448) ### [Apply Backpressure to AI API Workloads](https://matthewmodelflare.hashnode.dev/ai-api-backpressure?utm_source=hashnode&utm_medium=feed) Bounded in-flight work, deadline-aware queues, load shedding, fairness, and failure tests. 6 min read · [matthewmodelflare.hashnode.dev](https://matthewmodelflare.hashnode.dev/ai-api-backpressure?utm_source=hashnode&utm_medium=feed) [3](https://hashnode.com/posts/ai-api-backpressure/6ae3fd395597d31be248b448) [AumCreate](https://hashnode.com/@aumcreate) · [Oct 30](https://hashnode.com/posts/form-submissions-into-a-crm-without-losing-the-ones-that-fail/6ae417233055b2aa33ff4711) ### [Form submissions into a CRM without losing the ones that fail](https://aumcreate.hashnode.dev/form-submissions-into-a-crm-without-losing-the-ones-that-fail?utm_source=hashnode&utm_medium=feed) The thank-you message renders, the form clears, and the lead is gone. 8 min read · [aumcreate.hashnode.dev](https://aumcreate.hashnode.dev/form-submissions-into-a-crm-without-losing-the-ones-that-fail?utm_source=hashnode&utm_medium=feed)