All templates

Automated Typesense Index Sync Reconciliation

Every night, WebRun opens Typesense, compares each collection's document count against its source-of-truth record count, creates a Trello card when the two disagree by more than a small margin, and messages the on-call engineer on WhatsApp, so a stalled or failed reindex gets caught the same night instead of showing up as missing search results.

Runs on WebRun · Strict Lockdown policy
Every night at 11:30 PM WebRunorchestrates each step
1 Typesense compare index and source counts
2 Trello create the mismatch card
3 WhatsApp ping the on-call engineer
In short

How do I catch a Typesense index that has fallen out of sync?

WebRun compares every Typesense collection's document count against its source-of-truth record count each night, creating a Trello card and messaging the on-call engineer on WhatsApp whenever the two meaningfully disagree. This catches a stalled or failed reindex the same night it happens, instead of the team finding out later because search results were quietly missing records.

  • A stalled reindex gets caught the same night instead of surfacing as missing search results
  • The on-call engineer gets both counts up front instead of having to dig
  • Search stays trustworthy because it is checked every night

Built for Backend engineering teams · search platform teams · e-commerce engineering teams · data engineering teams

Step by step

What does WebRun do on every run?

The exact actions WebRun takes, in order - in plain language, so you can adjust anything.

  1. WebRun signs in and gets to work

    Opens cloud.typesense.org/login in a real browser with your saved login - no setup, no API keys.

  2. 1
    Typesense - compare index and source counts
    • Open Typesense and review the document count for each tracked collection
    • Compare it against the source-of-truth record count for that data
    • Flag any collection where the counts differ by more than a small margin

    Done when Every collection's document count is compared against its source with any mismatch flagged.

  3. 2
    Trello - create the mismatch card
    trello.com
    WebRun in Trello: create the mismatch card
    WebRun opens Trello to create the mismatch card.
    • Open the Search Index Health Trello board
    • Create a card for each mismatched collection with both counts
    • Place it in the To Investigate column

    Done when A Trello card exists for every collection currently out of sync with its source.

  4. 3
    WhatsApp - ping the on-call engineer
    whatsapp.com
    WebRun in WhatsApp: ping the on-call engineer
    WebRun opens WhatsApp to ping the on-call engineer.
    • Message the on-call engineer on WhatsApp
    • Include the collection name, the index count, and the source count
    • Ask them to check whether the reindex job needs to be rerun

    Done when The on-call engineer has been messaged about every index mismatch found tonight.

Run settings

How is each run configured?

Starting pageWhere Chrome opens at the start of each run
cloud.typesense.org/login
ScheduleRuns automatically on this cadence
Every night at 11:30 PM
DeliveryHow each run's result reaches you
Index sync alert · WhatsApp
OutputWhat each run produces - A nightly Trello card for any Typesense collection out of sync with its source, with a WhatsApp ping to the on-call engineer.
Alert
Setup & safety

Secure by default

Connect once, stays signed in

WebRun signs in once and keeps each session in a persistent environment, so every run picks up right where it left off.

Your credentials stay in your own private environment - WebRun never stores your passwords.
Strict Lockdown

Every action is checked against this policy before it runs.

Domains ALLOWLIST
Typed input ALLOW
Shell command BLOCK
File uploads BLOCK
Runs in a contained environment More on policies
Good to know

Questions, answered

Will WebRun trigger a reindex itself?

No. It only detects and reports the mismatch. Triggering a reindex or investigating the cause stays a decision for the on-call engineer.

What counts as a meaningful mismatch?

A difference beyond a small margin that could reasonably come from records changing during the check itself, so a one or two record gap from normal timing does not trigger a false alarm.

Does it check every collection?

Every collection you connect it to and give a source record count for, so the reconciliation covers your whole search index, not just one dataset.

Put this on autopilot.

Turn it on in minutes - or have our team set it up for you.