Monitoring IBM DB2 Server Version 8.0 (and above)

Figure 1 depicts the specialized DB2 UDB monitoring model that eG Enterprise offers for monitoring an IBM DB2 server ver. 8.0 (or above).

Figure 1 : The layer model of an IBM DB2 server version 8.0 (or above)

Every layer of Figure 1 is mapped to a wide variety of tests that a single eG agent executes on the DB2 server; these tests extract a wealth of performance metrics from the DB2 server.

One of the key qualities of this eG agent is that it is capable of measuring the internal health of the DB2 server from anywhere in the environment! In other words, the eG agent need not be present on the DB2 server to monitor its internal operations. The eG Enterprise system supports “agentless monitoring” of a DB2 server, by means of which, the agent can be installed on any remote host in your environment, and can be easily configured to pull out statistics of interest from within the DB2 server.

The statistics so collected enable administrators to find quick and accurate answers to the following performance queries:

  • Is the DB2 database server available? How quickly does it respond to user requests?
  • What is the current connection load on the DB2 database manager? How many of these connections are local, and how many are remote connections?
  • How quickly does the DB2 server process requests from client applications? Are there sufficient agents in the agent pool to service all the client requests?
  • Are the agents in the pool utilized optimally, or are too many agents idle?
  • Does the database server perform sorting efficiently? Has adequate sort heap space been allocated to the database manager to enable this?
  • Are sort overflows kept at a minimum?
  • Does sorting take too long?
  • Are lock escalations occurring too frequently on the database?
  • Are too many deadlocks been detected?
  • Do applications obtain locks quickly, or do they have to wait too long for locks?
  • Are the database buffer pools adequately sized?
  • Are the page cleaners and prefetchers been utilized effectively?
  • Are too many rollbacks happening on the database?
  • Have too many SQL statements failed?

The sections to come elaborate on each layer of Figure 1, the tests associated with them, and the statistics they extract.