You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The request_id added with #1650 is great for tracing vault requests and responses.
It would be a great to be able to trace a request among and across services including vault (and consul and nomad - hopefully).
With my internal services, I have some HTTP middleware that checks for the HTTP header X-Request-ID (the specific header is actually configurable) on each incoming request.
If the header is found and meets some arbitrary length constraints, each service uses the provided value as the request_id internally. This value is written to the request and system logs and we use it to lookup a complete request chain in our centralized logging systems.
If the header is not found on the request, a UUID is generated just like vault does today.
At least for my own services, this was a very simple but very powerful way to support distributed tracing and help in debugging issues.
I can't think of any security issues with a feature like this, but if there were concerns, it might be reasonable to allow a vault administrator to specify the name of the HTTP header as part of their listener configuration to enable this feature.
The text was updated successfully, but these errors were encountered:
You can already do this; in 0.6.5 you can set a list of headers that you want to appear in audit logs, including whether you want those values HMAC'd or not.
The request_id added with #1650 is great for tracing vault requests and responses.
It would be a great to be able to trace a request among and across services including vault (and consul and nomad - hopefully).
With my internal services, I have some HTTP middleware that checks for the HTTP header
X-Request-ID
(the specific header is actually configurable) on each incoming request.If the header is found and meets some arbitrary length constraints, each service uses the provided value as the
request_id
internally. This value is written to the request and system logs and we use it to lookup a complete request chain in our centralized logging systems.If the header is not found on the request, a UUID is generated just like vault does today.
At least for my own services, this was a very simple but very powerful way to support distributed tracing and help in debugging issues.
I can't think of any security issues with a feature like this, but if there were concerns, it might be reasonable to allow a vault administrator to specify the name of the HTTP header as part of their listener configuration to enable this feature.
The text was updated successfully, but these errors were encountered: