We use Find to a lot more then just search pages, and I guess that we are not alone doing this. For example gather information that should be shown on the start page. Getting related articles etc. But now and then Find becomes unavailable. How do you handle this on your sites? Why I'm asking is that we often recommend customers to use Find because it cuts development time (and it's a great search enginge). But creating backup solutions for handling down time from Find isn't very popular to the customer, because the perk of cutting development time is more or less gone. Would be interesting to hear if you have come up with some great idea how to solve this.
Do you have a fallback way of getting the information? Do you just hide it? Do you have a middle layer that Find updates and getting info from that layer?
Very good question and something I am also looking into, today it were a problem with the hosting so the question is very up to date!
I have no answer on it saidly
I would like to know the answer as well and I am also interested in how often it goes down.
I am thinking about something Ted Nyberg said the other day - that it could be used to partially replace caching. This would be suitable for my project as well that needs to go Azure with a pretty complex setup and will use Find anyways. However, if downtimes are often, then is this really a good idea?
Maybe http://status.episerver.com/ should be updated to include uptime stats.
An interesting and relevant topic. Normally, unless really business-critical, we simply hide the content - logic being we have the same up-time expectations for the Find service as we do for other EPiServer hosting services. However, I believe this was a bigger issue a couple of years ago before the Find hosting was scaling sufficiently. However, caching Find results is a fairly viable option I think for dealing with some fundamental use cases like a "most recent news" list etc.
Great idea Arild!
We have the same approach as Ted is mentioning and it is working ok as far as I can tell.
From our(EPiServers) perspective Teds approach is what we recommend. Caching is a really powerful way of both reducing roundtrips and also handling shorter outages. I also agree with what Ted said about uptime and us learning to scale in a secure way. We have learned a lot these almost 4 years that we have been running Find. Some perhaps interesting information about Find.
Today we run 12 Find Clusters! Located geographically distrbuted to SE, EU, US, AP.
Last week we handled 255,450,966 requests towards all our clusters combined. 255 Million! We provisioned a new cluster last week as well and will already this week setup yet another one.
We also doing like Ted says.
Cool numbers Marcus!