Sandbox Limiting
Oso Cloud developer accounts all run on shared infrastructure. As such, there are controls in place to ensure that all developer users have a good experience using the platform. Among these are limits on the number of concurrent requests for any Environment.
Types of limits
There are two different limits in place: one on the number of overall requests, and another on the number of write requests. Note that the latter is a subset of the former, meaning that write requests count toward both limits. As such the limit on overall requests is much higher than the limit on writes.
Overall requests
This is simply a limit on the number of requests a particular Environment will handle before it begins
to respond with 429 Too Many Requests.
Write requests
This limits the number of writes an Environment will handle before it begins to respond with
429 Too Many Requests. Fact addition and deletion are covered by this rate limit.
What to do if you've been rate limited
If you exceed either rate limit, you will begin to receive responses from Oso Cloud with the status code 429 Too Many Requests.
The easiest thing to do is reduce the volume of requests you are sending Oso Cloud and wait for the limit to expire! Rate limiting is very transient; it will expire in less than a minute and has no "memory" (i.e. you won't end up on some sort of list of users that are more likely to be rate limited in the future).
Obviously this isn't always possible, so other options include:
Use bulk_tell instead of tell
This is the most common cause of rate limiting. Thankfully it's also the easiest to fix.
Rather than distributing related facts across multiple requests, you can use
the client bulk methods to atomically write multiple, related facts (limit of
20 for Oso Cloud developer accounts) in a singular request. In addition to
transactional semantics, this has the benefit of reducing the number of write
requests you may need to send data to Oso Cloud.
Even though we are are writing three facts, this is just one write request:
await oso.batch((tx) => {
tx.insert(["has_role", {type: "User", id: "1"}, "member", {type: "Organization", id: "acme"}]);
tx.insert(["has_role", {type: "User", id: "2"}, "member", {type: "Organization", id: "acme"}]);
tx.insert(["has_role", {type: "User", id: "3"}, "member", {type: "Organization", id: "acme"}]);
});Talk to an Oso Engineer
Whether you want to pair on refactoring a policy or talk about caching strategies, there are lots of ways to reduce the number of requests your application sends Oso Cloud. Feel free to schedule a 1x1 with an Oso engineer. We're happy to help.