I love that the 1.5 version includes related kb articles in the notes section of each query in the content pack. I'd like to see this taken further. Add a link to the KB each log result from a search.
In our environment we have different needs for log retention. For example my vsphere logs can be aged after a month while my firewall logging should never be deleted.
so i would like to have a retention option per host or something.
An administrator may deem some messages undesirable, either specific logs produced by a source or a whole source. Such messages may result in CPU/disk resources being expended unnecessarily. Undesired log messages may result in Log Insight's data store being rotated more quickly than desired. In terms of licensing, one may wish to drop log messages from unlicensed sources, restricting sources from which log messages are... more »
I was wondering if you guys could make the data be zipped/dedupped during the night? If it already does dedup or something I guess its fine but it would seem that log insight takes a lot more space then our other logging system. We are trying to move everything to log insight but space constraints don't seem to be matching up.
Provide the ability to dump raw events more than 20K at a time. The user should have the ability to run an export query and retrieve the results. (eg. User provides a date range and requisite filters.)
From a customer "I noticed that Log Insight can only communicate with one instance of vROps. We currently have two instances of vROps because we exceeded the metrics allowed in the first instance. I’m not sure if this will be possible in future releases, but it would be beneficial to be able to communicate with more than one vROps instance."
Extend vR Ops to integration beyond notification and launch in context: deliver metrics (i.e. number of events, number of error / warning events, number of tasks, kernel warnings, VOB events) to vROps and match it to existing vROps objects (vCenter, hosts, VMs).
Dashboards make it possible to view data visually over time and alerts make it possible to get notified about events seen in the environment. What is needed is true reporting and more specifically: * Scheduling of when to run reports (both time of day and frequency) * Ability to at least email reports (pdf, html, and csv should be supported at a minimum) * Ability to include dashboards in reports (including legends)... more »
The agent should support globs (asterisk and wildcard) for folders. THe use case is IIS where multiple domains exist on the same server. Something like this
So then I could make one that does them all type thing.
Globs are supported for files so this is an inconsistency in the product as well.
Currently there is no alerting when Log Insight Master or Worker is not receiving logs/API events from its workers or agents. Part of the PCI compliance requires notification when a stoppage of logs is detected.
If this could be an alarm, or an automated email that is sent out, and have the ability to set certain thresholds (no logs within 30 minutes, 60 minutes, 3 hours, etc), that would be great.
Log Insight should facilitate understanding transactional flows, where a group of log messages tell a story together. The transaction identifier should be definable in content packs and by users, similar to an extracted field. For example: - vCenter, vpxa and hostd tasks are identified by an opID, relating task Start, Finish and subtasks within. - ESXi vMotion tasks are identified by an MigID, relating Source and Destination... more »
As a support engineer at a VMware partner, I would like to be able to leverage a support-centric version of Log Insight with the ability to quickly import customer support bundles and process them for diagnostic troubleshooting. I currently have a workflow where I do this, but it is quite tedious and only makes sense when I have a ton of logs/data to process.
Allow for the reception and logging to SNMP traps; also allow for sending SNMP traps as alerts. All configuration should be in the UI.
As I understand it - today LI saves data to the node's local storage(VMDK) and if that node is lost(deleted), the data is lost. It would be beneficial to (in cluster mode) be able to cluster the storage so in events of data lost. Data could be recovered from the other nodes in the cluster, by inserting a new node instead of the lost node and data would then be replicated to the node. I'm up for other ideas to the issue,... more »
It would be useful to be able to purge all events associated with a host when it is retired from service.