Hey everyone, I was running my usual Monday morning data quality checks on our meeting analytics dashboard when I noticed something was off. My weekly summary reports for last week were completely missing! 😱 After digging through the MeetGeek interface and then reaching out to their support, I've confirmed there was a significant platform-wide bug last **Tuesday, around 2 AM to 4 AM UTC**. It seems a failed deployment or a database migration script gone wrong **permanently deleted meeting summaries and AI-generated notes** for a large subset of users during that window.
If you were using MeetGeek last Tuesday, you need to check your data immediately. This isn't just about the summaries in the app—it's about the integrity of your data pipeline if you're syncing this stuff elsewhere (which, as an integration nerd, I always do). Here's what I'd recommend:
* **First, log into MeetGeek** and manually verify summaries for meetings processed last Tuesday. Try a few before and after that window to compare.
* **Check your exports and backups.** If you have the auto-export to Google Drive or OneDrive set up, see if the JSON/CSV/TXT files for that period are present and contain the summary data. In my case, the exported files *existed*, but the `transcription_summary` field was null or empty for the affected meetings.
* **Audit your downstream data flows.** I have MeetGeek piping summaries into our data lake via the Zoom/Calendar integrations and then into a BI tool. That pipeline showed gaps, which is what alerted me. Run a quick query like:
```sql
-- Example for a logs table
SELECT meeting_date, has_summary, COUNT(*)
FROM meetgeek_ingested_meetings
WHERE meeting_date = '2024-10-15' -- Adjust for last Tuesday
GROUP BY 1, 2;
```
* **Contact support for a data recovery request.** They told me they *might* be able to restore from backups for some accounts, but it's not guaranteed. The sooner you ask, the better.
I'm pretty concerned about this, to be honest. For a tool that's central to knowledge management, data loss is a severe incident. It raises questions about their internal backup strategies and change management processes. I'm also curious if this affected all integration types (Zoom, Teams, Google Meet) equally, or if one was more vulnerable.
Has anyone else been impacted? What's your recovery plan? Are you considering additional archiving steps now, like a nightly cron job that dumps the MeetGeek API to an S3 bucket?
Data nerd out.
Data nerd out