Mike Aracic

LogicMonitor Staff
  • Content Count

    23
  • Joined

  • Last visited

Community Reputation

0 Neutral

About Mike Aracic

  • Rank
    Community Whiz Kid
  1. The alert trigger intervals work for no data alerts, too. For a few reasons, it's not good practice to have a no data alert set on a datapoint that also has a value-based threshold. (For one, the message should definitely be different) Is that what you're talking about?
  2. I had an interesting case today: A user reported that a collector had completely stopped collecting data for all its hosts. He\'d verified that the collector\'s service account was good and had the necessary permissions. I checked one of his collection tasks, and it reported WMITask failed. (msg=winproxy return status=400 errmsg=IWbemLocator::ConnectServer:Error: 800706ba:The RPC server is unavailable.I ran a WMI query against the host manually and got the same result. (that the RPC server was unavailable) As mentioned in our helpful article on WMI Troubleshooting, generally when the error is RPC server unavailable, the issue is one of basic connectivity, usually caused by a network or firewall issue blocking the WMI traffic from reaching its destination. In this case however, the user had the service account set up correctly, but the hosts had wmi.user and wmi.pass properties overriding those settings. This worked until the password for the user specified in those parameters expired, at which point all the hosts so affected ceased responding to WMI queries from the collector. Once the host parameters were removed, collection resumed immediately.
  3. If youd like to be able to support 2 distinct databases on the same host, its possible to create custom properties (which could be applied at the portal, group, or host level to the desired hosts) called: jdbc.mysql.dbname and jdbc.postgres.dbnamen nThen clone your existing postgres and mysql datasources and do the following to the clones: add && hasproperty(jdbc.mysql.dbname) or && hasproperty (jdbc.postgres.dbname) to the Applies to field and replace the DBNAME tokens with jdbc.mysql.dbname or jdbc.postgres.dbnamen nThis will associate the cloned datasources only with hosts that have the proper custom properties and ensure make those cloned datasources use the new tokens. Let us know if you need help with this.
  4. If youd like to be able to support 2 distinct databases on the same host, its possible to create custom properties (which could be applied at the portal, group, or host level to the desired hosts) called: jdbc.mysql.dbname and jdbc.postgres.dbnamen nThen clone your existing postgres and mysql datasources and do the following to the clones: add && hasproperty(jdbc.mysql.dbname) or && hasproperty (jdbc.postgres.dbname) to the Applies to field and replace the DBNAME tokens with jdbc.mysql.dbname or jdbc.postgres.dbnamen nThis way the clones of the datasources will associate with the hosts, and the datasources will use the application-specific custom substitution tags. Let me know if you need more information.
  5. There are two kinds of MySQL Slave monitoring. (Both described here) The first, basic kind, is slave status, which is an indication of how effectively the slave database is applying the replication information to itself. The second is replication lag, which uses a pair of datasources, along with a dedicated LogicMonitor database on the servers, to measure the actual amount of time between inserting a record in the master and being able to retrieve that record on the monitored slave instance. For that youll need to configure the credentials and add the database as described here.
  6. This looks like an issue with integrated security authentication. Could you contact us either at support@logicmonitor.com or using the Chat with Engineer button in your portal. Wed like to take a look at the configuration.
  7. LogicMonitor uses an ESX api to retrieve performance information from the ESX and vCenter servers. Weve noticed that, on some occasions, vCenter servers will return -1 for all counters from an attached vCenter host. Were still looking into the cause, but we have yet to see a case where an ESX host monitored directly did this. Because of this, and the ability to do hardware health monitoring, we suggest adding ESX hosts directly to monitoring. Send us an email with specifics or contact us using Engineer Chat, and wed be happy to look into your specific case.
  8. The issue was resolved once the user uninstalled the security patch described here. Its not clear whether running without said patch in place represents a significant risk. It may be possible to further patch SQL Server or change the JVM to avoid this issue. See article )
  9. I would advise anyone who has a device theyd like monitored in this way to do the following: Clone the appropriate datasource, SNMPIF_64- generally, and modify their copy as above, with no filter on UP and persistent checked, then apply this directly to the devices using the applies-to field in the datasource. Some help is here, and wed be happy to answer questions about the process.
  10. In order to do this, youll follow the directions on our script collector help page. Bear in mind that the script will run on the agent host. Its possible to associate the datasource with another host and pass that hostname as an argument to the script in order to reference, and associate results, with hosts other than the agent, but script must be written to run on the agent host.n nGiven that you have working scripts, you may prefer to install an agent on the machine that currently runs the scripts.
  11. Improving the way agents are monitored and alerted on is already topic on our development roadmap, and I will forward your feedback on to our product group.n nIn the meantime, the agent down alerts are fairly important alerts; when you see those, it means that no information has come in for the affected hosts for a significant interval. Its possible that nothing (other than the lack of data collection) may be wrong, but at that point, the affected hosts are effectively not under monitoring. The web services monitors on the services tab can be an effective backup, as they can be configured to test your outward facing services from multiple points of presence on the webb.
  12. Brian, if its convenient, grab us in engineer chat, or send us an email to support@logicmonitor.com and describe what hosts in the portal we can use as an example, and provide us with any information you have about how the data may be exposed.
  13. Our PostgreSQL replication datasource, with implementation described here includes some scripts used to extend snmpd. One of them uses date commands with a -u option. This assumes that your databases use UTC. If they do not report the time in UTC, this will cause a consistent error. Depending on the specifics, you may want to remove the -u, or possibly apply an offset in seconds to correct for the discrepancy.
  14. James,n nWe do not support a host property to allow for monitoring of an out-of-band IPMI source. There do appear to be options on both Linux and Windows to allow for pass-through of IPMI functionality to the in-band address, but youre probably already aware of them. They also require software to be running on the host. Let us know if you have more questions about this.n nThanks,n nMike
  15. I just clicked through it. Were you able to find what you needed?