Operators (support folks doing application monitoring) deserve all the simplicity they can get, sure. Administrators deserve some attention too. We're delivering on that with administrative dashboard improvements and more in-context administration. As usual, though, once one problem is put to rest, another shows up. The next target for simplicity is likely distributed administration.
Most Foglight customers run multiple Foglight servers - sometimes federated, sometimes multiple sites. But the bottom line is that they have several of these - sometimes dozens or more. At that point the administration problem moves beyond day-to-day operations using standalone UIs. A new set of issues presents itself:
- Consistency: How do we make sure the Foglight servers are all doing the same things the same way?
- How do we report on configuration?
- How do we apply changes easily?
- How do we move customizations, especially dashboards and reports?
- Audit? Cross-org audit?
- Mobility: Many customers want to load balance their child servers. Others need to make changes as a result of organizational shifts. How do we change where something is done?
- Moving agents between agent managers
- Moving agents between Foglight management servers
- Moving objects
- Deleting things that are stale
- Tuning: Beyond consistency, how do we make sure that all the servers are tuned and operating properly?
- Performance architecture templates
- Standard configuration for JVM, database, network
- Deploying resource monitoring for Foglight
- Federation: The need for federation alternatives is growing
- Federate with partial data streams from children
- Federation of federation
- Federation HA
- Temporary Federation instances, which implies quick spin-up/tear-down
- Federation access to children for administration
While there are solutions in place today for most of these problems, the solutions could be improved. I'd love to hear more about your distributed administration issues.