{
  "id": 14624814,
  "title": "Our standby received every change for five hours and applied none of them",
  "url": "https://urgent.news/2026/10/15/our-standby-received-every-change-for-five-hours-and-applied-none-of",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-10-15T05:55:24.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/sergey_shinder_ab2d943365/our-standby-received-every-change-for-five-hours-and-applied-none-of-them-41nh"
  },
  "original_language": "en",
  "account": null,
  "summary": "In August, a two-minute switchover of the main Postgres database was planned to patch the primary kernel. However, the standby, which serves reports for the finance team, received every change for five hours and applied none of them. This resulted in the standby being five hours behind in replay, causing services that write data to be down for forty-one minutes. The issue arose because when the primary vacuums away rows that a long query on the standby can still see, replaying that change would pull the rows out from under the query. Finance had previously set the max_standby_streaming_delay value to minus one, causing the standby to wait forever. To prevent this from happening again, the standby was given a longer delay of thirty seconds and a lag alert was implemented to measure replay. Additionally, a switchover was conducted in a staging environment every month to ensure that the standby can take over from the primary.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}