Emergency help

If one of the sites has gone down completely then everyone at Semantic will get an alert through to their phones. It would be picked up during waking hours UK time.

Here's the process we follow for Semantic sites. For legacy sites, we would alert the SJ team through Slack/email and escalate as needed.

1. Alert comes through

We'll receive the push notification from Statuscake (the tool we use to monitor uptime)

2. Try to replicate the error

We try visiting the site to determine if it was just a blip

3. Wait a few minutes for restart

If it's a blank white screen, then it's likely to be a site restart or hosting issue. If the site is not back within 10 minutes, we'd check the logs to confirm this. If it comes back, then the restart was successful, so no further action is needed. If not, then we continue...

4. Check logs and attempt a restart

Attempt to force a manual restart through the Azure admin area (this is often sufficient to bring it back).

5. Escalate

By this point, we should have an an idea of the cause of the error from the logs. We would then continue to bugfix within the CMS content, republishing or tweaking settings as needed. For larger issues, then more detailed code analysis would be carried out, with hotfixes deployed as soon as possible.