Is there any benefit to Argo & edge cache if using workers sites?

I don’t have exact statistics, but geographic distance tends to have more impact on Argo (on average) at least in my experience. So on an average request a user in Seattle with an origin in Houston will be more likely to be smart routed than the same user hitting an origin in LA and when you roll in international the % smart routed and percentage performance improvement tends to go up. But as a huge disclaimer, my experience is primarily with Cloudflare enterprise customers who often don’t have the same poor/underpowered origin issues that some of our pay-as-you-go customers might have.

That was a really long way of saying ‘it depends’.

In terms of overall speed I think generally:

Origin w/o Argo < Origin w/ Argo < KV < Object already in cache at edge colo

For me I wouldn’t enable Argo for Worker Sites, the KV fetch, even if remote is extremely comparable to a remote cache fetch and the price you pay for won’t be worth the few milliseconds you earn in potential network latency

Are you possibly suggesting our origins suck? :triumph: :wink:

1 Like

I generally think we are very much in theory-land at this point :slight_smile:

1 Like

Great discussion, guys, thanks!

Do you have a source for that? The workers sites documentation implies it’s always stored on all edge locations.

Workers KV read performance is determined by the amount of read-volume a given key receives. Maximum performance for a key is not reached unless that key is being read at least a couple times per minute in any given data center.

Obviously it could simply mean it’s not being cached locally, but I was told once (no clue how to fin that…) that there was a central store.

edit also @cloonan there is an error in the menu on the left in those pages, there is double “Limits” instead of “Pricing”

1 Like

Appears to be corrected…

1 Like


1 Like

This topic was automatically closed after 30 days. New replies are no longer allowed.