This page is an area to house relevant links and documents and to facilitate discussions around directions for instrumentation in the TIER releases.
TIER Instrumentation Phase 1 is currently being implemented according to this specification.
Questions for the next phase of instrumentation:
- Who is the audience for the information? TIER management? Deployers? Both?
- What data should be collected in the next phase for instrumentation?
- Both Grouper and Shibboleth can send rich sets of data to a remote collector
- There are several operational uses for this data
- Who's running version X of component Y?
- Who's running on OS Z?
- Who's running near RAM capacity?
- What IdPs haven't refreshed InCommon metadata in the last 24 hours?
- What Grouper deployments have not applied patch A?
- Should the collected data be constrained to 'for-public-consumption' classes of data?
- If not, how do we apply appropriate controls to the data to ensure it is handled appropriately?
- Are there "non-default" metrics required from Shibboleth that TIER needs to configure?
TIER recommendation for data privacy
TIER would like to expose data from the central collector in an anonymized public way. The institution name will not be tied to its data. But if the institution is identified in the instrumentation submission, TIER can contact the institution for example if there are patches that need to be applied.
Grouper instrumentation wiki. Grouper will be sending a format like this: