dc: alter active tracking to longer duration

[Mark-Simulacrum]

Security issue notifications

If you discover a potential security issue in s2n-quic we ask that you notify
AWS Security via our vulnerability reporting page. Please do not create a public github issue.

Problem:

Today the cleaner (https://github.com/aws/s2n-quic/blob/main/dc/s2n-quic-dc/src/path/secret/map/cleaner.rs) tracks "active" entries and reports counts for those entries. We clear the active status on each iteration, i.e., once a minute. We should bump up the timeline or extend active with multiple timelines (e.g., active in last 1m, 5m, 30m, 1hr)... but we need to be careful on memory usage and performance of the extra tracking so needs some care to get that right.

Need By Date:

n/a

Solution:

• Does this change what s2n-quic sends over the wire? no
• Does this change any public APIs? only new metrics

[maddeleine]

Hi are you saying that clearing the active status each minute is losing information/necessary detail and therefore we should try extending this tracking to longer timeframes? Just to get a more accurate picture of how frequently the entries are used?

[Mark-Simulacrum]

I think yes -- essentially, tracking just the last minute is too aggressive. The goal of this metric when added was to give a sense of what % of the map is actually in "active use", i.e., where we would expect pruning to start being a direct contributor to high handshake volume. The goal is to give a better metric than the total usage to track for customers wanting to know whether they are approaching map size limits.

I suspect extending the active period to 10-30 minutes might be enough for that.

[Mark-Simulacrum]

Let's also add a metric tracking usage based on entries that were created in the last 24 hours (the re-handshake period). That should help get a more reliable metric than the total usage given client socket address churn.