Handle Degradation of service?

Hey Algolia, neat library.

We’re trying to launch this in production, but we’re stuck at degradation of service. Let’s say we forget to update our billing and you shut us off, or your servers go down and the search doesn’t return.

How do we gracefully handle occurrences such as this? We want to be able to revert to our old flow without Algolia places in the event that your service cannot be reached.

Is there some kind of PING we can ping to your servers to check for service?

Or does the

places({ ... })

init have any error condition / catch?

[raphi]

@lightninglu10 you could use a proxy to intercept the requests you want or you can just edit your /etc/hosts file and redirect the domain to your localhost. The first request will fail (your localhost returning nothing) and it will retry on another domain.

For example: /etc/hosts

127.0.0.1       places-dsn.algolia.net

You will then see requests hitting another domain (we have redundancy on all our DNS domains in case the DNS fails… rare but it happened!).

Then, we have redundancy on servers as well. All our clusters at Algolia are redundant with 3 servers behind, placed in different datacenters. If one fails or answers too slowly, the client will switch to another one. On top of that, for Places, we have setup DSN servers all around the world, to be closer to the end user.

Let me know if that helps!

[Haroenv]

You can enable “offline” mode in your devtools