Forum Discussion

Anonymous's avatar
4 years ago

2021-03-31 Office Hours

We wanted to call special attention to this month's last-Wednesday webinar as it will be slightly different than our normal Office Hours. We will still have a Q&A section, but before we get to the questions, we'll have @Sarah Terry, Sr. Director, Product Management on the webinar to present a new segment called "What's new at LM?" She'll be talking about LM Logs, our new log analytics feature. Register here then bring your coffee and questions!

1 Reply

Replies have been turned off for this discussion
  • Anonymous's avatar

    Thanks to everyone who attended today's webinar. Please consider leaving us feedback here. Your support keeps this program improving!

    a href="" rel="">You can find details about our Trace beta program and register to join here.

    Here's the recording from today's webinar:


    Here are the questions asked and where you can find the question asked and answered in the video:

    [12:09] Any thought to bringing logs into dashboard widgets for visualizing patterns?

    [12:57] Periodically, its not uncommon to enable debug/more explicit logging for resources....suddenly increasing volume might trigger sudden false positive anomalies.

    [14:11] any thoughts to allowing some kind of SDT for logs?

    [14:44] And, is there a 'negative' anomaly, when a very chatty resource suddenly goes radio silent?

    [18:12] Would you share some thoughts about designing an snmp datasource to capture a counter in a Juniper firewall and convert it to a rate? 
    [24:10] jnxJsScreenMonEntry

    [26:38] Since I'll be polling about a dozen of those counters, that means a mult-instance datasource (probably the better way than having a datasource for each individual counter).   I can just repeat what you just described---that was a single instance I think--- for each wildvalue/instance?  Thanks for letting me monopolize the conversation!

    [32:41] As you noted that is a long list.   we are currently only enforcing a small subset, so only those counters will be non zero and increasing.  Is there a good way to allow it to walk the MIB and dynamically pick up and record only those that are non-zero?   That way if/when we add enforcement to counters in the future I wouldn't need to return to the datasource to add a new datapoint?