Another day, another UI "improvement" that breaks core functionality. ContentBot pushed an update last week and now half my saved templates are just gone. Poof.
No warning, no export option offered before the change. Just logged in to a fresh new look and a bunch of empty slots where my carefully crafted templates used to live. Checked the usual suspects: browser cache, different machine, incognito. They're not there. The API still shows my old template IDs, but they return 404s. So much for backwards compatibility.
I'm back to storing templates in a git repo and generating them with a simple script. At least version control doesn't randomly discard my work. Here's the hack I'm using now to avoid their UI entirely:
```bash
#!/bin/bash
# Store template in a file, push via API
CONTENTBOT_API_KEY="..."
TEMPLATE_NAME="$(jq -r '.name' template.json)"
curl -X POST "https://api.contentbot.example/templates"
-H "Authorization: Bearer $CONTENTBOT_API_KEY"
-H "Content-Type: application/json"
-d "@template.json"
```
Anyone else find their templates missing, or are we just trusting the cloud to hold our stuff now?
If it ain't broke, don't 'upgrade' it.
What you're describing is a critical failure in their data migration and versioning strategy. The fact that the old template IDs persist in the API but return 404s is particularly telling; it suggests a hard deletion of the underlying records or a broken foreign key relationship in their database schema after the UI update.
Your move to version-controlled templates is the correct architectural response. I've implemented similar safeguards for client Salesforce deployments, where any complex report, dashboard, or email template is treated as configuration-as-code. The core principle is that the system of record should never be solely the vendor's UI. A scripted deployment pipeline, triggered from a repository, becomes mandatory for business continuity, not just a nice-to-have.
This incident highlights a broader pattern in SaaS tooling where UI updates are treated as purely front-end changes, with inadequate back-end impact analysis. Has ContentBot's status page acknowledged this as a known issue, or is support blaming local cache?
measure what matters