This solves two problems:
1. redundant trackers making redundant requests
2. trackers being obliterated as soon as the primary metric component
is destroyed
It introduces a new problem where visiting more and more node and
allocation pages adds to an ever-growing list of trackers that can
assume lots of memory, but it solves the problem by using a
least-recently-used cache to limit the number of trackers tracked.
The lowest valid blocking query index is 1, but the API will return 0 if
there has yet to be an index set (no response). This in conjunction with
that 0 being stored as a string made the "fallback to 1" guard not work.
The UI will no longer try to redirect to the appropriate namespace or
region if one is found in localStorage. Instead, it will assume that
the lack of query param means the default namespaces or region is
desired.
This is incredibly tricky with query params, since there is a bundle of
timing issues, lifecycle issues, missing features, and all around
gotchas with query params.
This solution has no observers and no instances of the system service
being set from the jobs controller.
The upside to this is no observers, much easier to follow logic, no more
dependent key chain reactions.
In the event that a namespace in localStorage is not found in the
response for /namespaces, clear the localStorage value and assume
the default namespace.
- Service to manage X-Nomad-Index values
- Adapter method for reloading relationships with additional params
- Pattern for watching models and model relationships using EC
Michael Lange