Ops Tools
465 results found
-
Add specific label to Arbiter deployment to distinguish it label-wise
As per https://github.com/mongodb/mongodb-kubernetes-operator/issues/1442 we would like to request to either and/or:
a) have the operator add a new label to replica and arbiter statefulset deployments to distinguish them from each other
b) have the ability to add your own labels to specifically the replica OR the arbiter statefulset deploymentsHaving the ability to easily filter arbiter statefulsets from non-arbiter statefulsets would be great. This is useful for example for backup tools that run commands on pods based on a combination of labels, to exclude any arbiters from backup.
3 votes -
Operator crashes when it doesn't have permissions to watch a namespace
If one of many namespaces does not set the permissions for the operator to watch the namespace, it throws exceptions and goes to crashLoopBackOff state.
This is clearly a bug. One misconfigured namespace should never be able to take the operator down with it.1 voteThis is expected behavior and common among operators; it can't function and do what's needed if it lacks the permissions needed.
I know we have an open support case around this to try and understand more about your use case, and we're hoping that we'll be able to offer some guidance to avoid this problem and still achieve what you need. It may even be a new use case that we look to support.
-
Identify the snapshot of each project and its size in S3 blockstore
Calculating the storage size consumed by snapshots for each projects deployments for the storage size consumed individually in S3 Blockstore. However, on the S3 snapshot store the data will be stored in the below format which doesn't include any project id to identify the specific project.
s3://bucket_name/0E3AA1971D5CF1CA52F9AF22A4228F10293AE9804D43FBF7EB5DDE38DB06B74A/5b27b0e4083826088f259f28_A s3://bucket_name/1860B12165FB7ED336DDAB9D306EF38E18FCBD36BF695904C497B825F83581DC/5b27b0e4083826088f259f28_A
This feature is helpful for the customers to understand and segregate the storage consumption by each deployments4 votes -
Allow assigning users to an organization via api by adding a bypass parameter
Via the parameter
bypassInviteForExistingUsers
it is possible to a add a user to a project.
It is important to also have this feature for the Organization, so please add a parameterbypassOrgInviteForExistingUser
or similar to enable this.
Without it, we can not automate the creation of new environments.1 vote -
Allow configuring external applicationDatabase running on MongoDB Atlas
Allow configuring external applicationDatabase running outside of Kubernetes in the opsManager configuration.
1 vote -
Management capabilities for trusted CA certificates in the Admin UI
Currently it is a manual process per Ops Manager server to add trusted CA certificates to the Java TrustStore. This store is also overwritten on Ops Manager upgrades.
Similar to the S3 block store option to use a specific CA, it would be useful to manage trusted CA certificates for other endpoints such as HTTPS proxies, webhook alerts, and custom download URLs.
1 vote -
Need an alert for KMIP master key rotation
Currently opsmanager has manual rotation of KMIP master keys.
Enhancement required:
Need an alert from project level to rotate keys when nearing the schedule rotation just like SSL expiration alertsNeed for automation of this key rotation would be helpful. If fails, it needs to revert back changes and alert for human intervention.
1 vote -
Allow service spec changes via MongoDBCommunity CR
I aim to include extra sidecar containers, such as a metrics sidecar, allowing me to deploy an external metrics exporter alongside that would be scraped via service. I intend to configure annotations on the service for Prometheus automatic discovery, enabling the scraping of metrics and exposing the exporter container at the service level, similar to the one exposed when prometheus integration is enabled.
Updating the service spec is not currently possible, and as result, a lot of manual steps are required, like creating a custom service for each cluster, which points to the sidecar container with the required annotations in…
1 vote -
Add support to convert a replica set to a sharded cluster in Kubernetes
Currently it is not possible to use Ops Manager UI or Kubernetes Operator to convert Replica Set to Sharded cluster. Error : changed not allowed.
2 votesPlanned - we have a requirement to have this completed in time for MongoDB 9.0. But no timeline (yet) for delivering this within that timeframe.
-
disable auth on metrics
Opentelemtry-Collector does not support secrets for ServiceMonitor/PodMonitor resources, which generates authorization issues while trying to scrape the metrics endpoint of MongoDB.
I'm looking for a way to disable the basic_auth on the metrics endpoint of MongoDB, I already tried a lot of ways, including an empty username/password, but nothing worked, any help would be highly appreciated.
1 vote -
Add Public API Endpoint for Backup Snapshot Status
When a snapshot is ongoing, the only location to find information on the current status is in the UI on the backup page for a project.
An API endpoint with the same information as the UI would allow for automation and alerting tools if the snapshot creation is not performing as expected.
2 votes -
Create an option to turn ON/OFF daily snapshots schedule from "Ops Manager- Continuous Backup - Edit snapshot schedule"
Currently, we have the option to take daily snapshots at different intervals (6 hrs, 8 hrs, 12, hrs, 24 hrs) but we do not have an option to not configure daily snapshots. While daily snapshots are considered essential, we have almost 200 replica sets of different DB sizes and the daily snapshots, tend interfere with itself due to large DB size and not completing itself in good time for the next day's daily snapshot to run. So if we turn off the daily snapshot or tune the frequency of the daily snapshots on select large replica sets, we will avoid…
10 votes -
Add comprehensive configuration options for persistent volumes
From a project point of view, we need additional configuration options for persistent volumes. For instance, as far as we know, each MongoDB Deployment that uses a persistent volume is configured with Access Mode ReadWriteOnce per default. We need to customise this option for optimal usage. For configuration options would be a plus.
1 vote -
Add support for zstd compression in Ops Manager connection string
I just tried to enable zstd on the
mongo.mongoUri
connection string for Ops Manager and it looks like it's not supportedFeb 6 13:53:20 opsmanager1 mongodb-mms[3971104]: An unexpected error occurred during pre-flight checks: com/github/luben/zstd/ZstdInputStream
Feb 6 13:53:20 opsmanager1 mongodb-mms[3971104]: java.lang.NoClassDefFoundError: com/github/luben/zstd/ZstdInputStreamGiven the high network in/out values we see from the mongo underlying Ops Manager, I think zstd is likely to be very effective and helpful!
2 votes -
SAML sign either Response or Assertion in Ops Manager
Ops Manager currently requires both the Response and Assertion to be signed for SAML auth. Our IDMS system cannot authenticate via SAML as it can sign either one or the other, not both. We would like a configurable option in Ops Manager SAML auth to enforce signing of Response or Assertion.
1 vote -
Allow conditions for all alerts based on DB/cluster name
We have a need to route alerts based on our database/cluster's impacted instead of having them apply to any in the current project. I know some of them can condition based on host name, which uses the cluster name, but most alert types have no similar option.
Examples would be:
Search Index Build Complete
Search Index Build FailedThese are sent without conditions and even the message does not indicate what cluster they are for (although you can guess based on the collection name provided).
We do not have the option to route these ourselves with a webhook, as the…
7 votes -
mongod startupWarnings
Create an "alert" to send notificactions when a mongod proccess has, for any reason, startup warnings.
e.g.
1)
The configured WiredTiger cache size is more than 80% of available RAM. See http://dochub.mongodb.org/core/faq-memory-diagnostics-wt2)
/sys/kernel/mm/transparent_hugepage/defrag is 'always'.3) Others.
1 vote -
Change the "Replica set has a late snapshot" Alert from Global to Project level
Currently the "Replica set has a late snapshot" Alert is a Global Alert. It would be useful to have this changed to a Project level Alert so that the Alert can be tuned for each specific deployment to provide better customization of the Alert.
2 votes -
Disable point-in-time restores
It would be nice to have the ability to set the parameter "Allow point-in-time restores going back" to ZERO (disabling PIT restores). This could be useful in situation where a database is producing a lot of oplog and DBA wants to avoid the saturation of oplog-store. In other words: "I want to mantain shapshot backup functionality, but deactivate PIT functionality".
1 vote -
Incremental backup with reduced snapshots size
The actual size of EACH snapshot is equal to datapath filesystem size. Every snapshot is essentially a copy of all db files. This prevents the use of OPS Manager Backup for large databases. It would be desirable to have a "real" incremental backup, based on one "baseline-snapshot" plus "delta-snapshots" of reduced size.
1 vote
- Don't see your idea?