mirror of
https://github.com/hashicorp/vault.git
synced 2025-08-16 03:27:01 +02:00
a2de4c75cd
Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Brian Shumate <brianshumate@users.noreply.github.com>
72 lines
3.0 KiB
Plaintext
72 lines
3.0 KiB
Plaintext
### Can I view the list of unique client IDs that contributed to my client count aggregate? ((#view-data))
|
|
|
|
**Yes, if you are using Vault 1.11 or later**.
|
|
|
|
As of Vault 1.11, you can export the list of unique
|
|
clients that contributed to your client count for a given billing period with
|
|
the
|
|
[activity export API endpoint](/vault/api-docs/system/internal-counters#activity-export).
|
|
|
|
### Is clientID viewable in the audit logs for non-entity tokens? ((#clientid-in-logs))
|
|
|
|
**Yes, for Vault v1.9+**.
|
|
|
|
As of Vault v1.9, audit logs include a field called `clientID` that records the
|
|
active or computed entity ID of the associated token.
|
|
|
|
### Can I skip client computation for a period of time during the billing period? ((#skip-computation))
|
|
|
|
**Yes, but the data may be inaccurate**.
|
|
|
|
Breaking up the data will likely result in client duplication. For example,
|
|
assume your billing period runs from January 1st to December 31st, and you break
|
|
the computation into two periods to exclude the month of April:
|
|
|
|
- Period<sub>1</sub> runs from January 1st to March 31st
|
|
- Period<sub>2</sub> runs from June 1st to December 31st
|
|
|
|
Vault treats the two requests independently so January 1st and June 1st are both
|
|
used as a fresh start to re-count unique clients. Any client that was active
|
|
during both periods will appear in both result sets, even though Vault will only
|
|
counted that client **once** for the actual billing period.
|
|
|
|
### Are there conditions that will cause me to lose client data? ((#lost-data))
|
|
|
|
**Yes**.
|
|
|
|
The Vault activity log handles client computation. The standby nodes track and
|
|
accumulate activity log data then transmit updates to the active node over gRPC
|
|
whenever the log grows by 8KB or 10 minutes has elapsed, whichever happens
|
|
first.
|
|
|
|
If a node goes down during the accumulation phase you will lose the
|
|
untransmitted data in addition to any client count data that would have been
|
|
collected during the outage.
|
|
|
|
### What happens if a portion of the data is missing for my billing period? ((#missing-data))
|
|
|
|
**Vault only returns the most recent contiguous set of data**.
|
|
|
|
For example, assume your billing period runs from January 1st to December 31st
|
|
but you disabled tracking for the month of April. Vault will look for the
|
|
largest, contiguous window of time where data is available, May through December,
|
|
and compute results for that period of time. May 2021 through December 2021.
|
|
|
|
### Can I disable client counting? ((#disable-client-count))
|
|
|
|
**Yes**.
|
|
|
|
You can use the Vault API to
|
|
[update the client count configuration](/vault/api-docs/system/internal-counters#update-the-client-count-configuration)
|
|
and disable the tracking parameter. If you disable client counting in the middle
|
|
of a month, Vault will discard any data currently recorded for the month. Data
|
|
for previous months is preserved.
|
|
|
|
### Can I configure Vault to log the client count data? ((#log-client-count))
|
|
|
|
**Yes**.
|
|
|
|
You can use the Vault API to
|
|
[update the client count configuration](/vault/api-docs/system/internal-counters#update-the-client-count-configuration)
|
|
and specify your preferred retention period.
|