Downtime June 24th 2019

Hello everyone,

we are very sorry for the outages that you have reported, the summary is that two outages happened together impacting different areas of DatoCMS.

One was Cloudflare: which would have impacted the GraphQL endpoint which had given time outs to certain IP ranges.

The other was with Heroku, which have impacted both GraphQL and REST APIs:

Now everything should be back to normal, but please let us know if you are still experiencing any issue.

We are very sorry for this.

A followup article from Cloudflare:

which gives a lot more information on what went wrong.

I was asked to bring my Slack conversation over to this site instead. I don’t want to copy and paste everything here though as it seems redundant. The relevant links are:

The short version is that I expressed disappointment about the fact that DatoCMS didn’t update its status page with information about the outage and didn’t communicate clearly about what was going on. There were some challenges (it was a holiday, the status page was on Netlify and couldn’t get updated, etc.) that made this outage somewhat unique.

Where the conversation was left was that there was some need to update the status page to make it clearer what service providers DatoCMS is using so that customers don’t have to do their own digging to figure out that if CloudFlare has an issue, that it might be impacting DatoCMS because DatoCMS uses CloudFlare.

This was the final comment in the thread before I was asked to move the conversation here:

As long as we’re talking about this, what services are you using that we should be aware of?

• CloudFlare clearly
• Heroku? (I thought maybe you just had a lot of Heroku customers using DatoCMS, but maybe I was misreading it and you use Heroku as well)
• AWS?
• Netlify for anything other than the status page?

What should we keep an eye on?

I would still love answers to the question of what services we should be aware of until the status page has been updated to reflect the services that DatoCMS relies on.

Hello @grigs thank you very much for taking the time of writing here.

Here’s a list of status pages of the services that we are using ourselves, that we’ll integrate in our status page: