Technical Tips & Tricks
Previously internal product knowledge now available for you to explore!
- 1,744 Topics
- 373 Replies
Last tested: Aug 6, 2018 If a user does not have the `see_drill_overlay` permission in their permission set, they will not be able to view links off fields (i.e. links defined by using a link: parameter). This content is subject to limited support.
What does a history status of 'hold' mean?Knowledge Drop
Last tested: Dec 18, 2020 It means another query with the same render key (exact same query) is currently running. We use this status to avoid having duplicate queries run against the db, and instead put one on hold so that it can pull from cache after the query successfully runs. This content is subject to limited support.
Last tested: Jul 16, 2020 It is likely our feature, In-App Guides. This feature (a 3rd party feature by Pendo) provides helpful guides that walk you through parts of the product, introduce webinars, etc.Why aren't all my users seeing it?Pendo administrators generally ramp up the guides, starting by targeting a random 0.5% of all Looker users, and increase them over time. For major announcements, like chat support on holiday, the guides will be shown to all users that qualify ( in that example, all admins and developers).Can I turn it off?Yes, an admin can turn it off in the Admin Panel. The button looks like this. This content is subject to limited support.
Does Looker perform Git garbage collection?Knowledge Drop
Last tested: Aug 6, 2018 Looker performs garbage collection only on startup, and only if garbage collection hasn't been done on the given repo for over 30 days.It includes both prod and dev folders. It's smart enough to perform this in the shared filesystem in the case of a cluster. This content is subject to limited support.
Last tested: Mar 31, 2020 Accidents happen. This one's happened a few times: A user has a project A connected to a git repo, they create a new project B and mistakenly link it to the same remote git repo. How can they split these projects back up? Unfortunately the fix here will be a bit manual.First thing is to make sure you have a second remote repository ready for project B. Point project B at the new repo - if the code from project A has already been synced into project B from the mistakenly shared remote repo, then you will need to manually delete the files from project A. Then, push project B to remote - this will push to the new project. Now, you need to get project B files out of project A. Manually delete these from project A and push these changes - this will push to the original git repo. This content is subject to limited support.
Last tested: Jul 4, 2018 This error implies a network timeout. Check for a load balancer on your end. On Looker nginx servers sample we use a timeout of 3600 seconds, which can be customisable to your environment. This content is subject to limited support.
Interactive API docs page UnresponsiveKnowledge Drop
Last tested: Jun 7, 2018 Try using the endpoint with a response format that has less data (if possible). The API call is probably sending back more data than the browser can handle. The Interactive API Docs page is nice for super lightweight testing but isn’t meant to process much data.You can also try running this from the command line to see if it works. This content is subject to limited support.
PDT’s won't work with Heroku?Knowledge Drop
Last tested: Nov 16, 2017 Not exactly, but there are workarounds.The problem is that Heroku only offers a streaming replica Postgres DB that is a read-only slave. This is because of Heroku's "hot swap" architecture, which basically means that the DB can swap out and become the production DB under certain circumstances, and therefore we cannot write to it.WorkaroundsYou can use a forked database and use the Update Connection API endpoint to change the connection credentials programmatically, but that would be a unique approach. Here’s a doc for that:https://docs.looker.com/setup-and-management/database-config/postgresql-herokuHowever, the best option here is to setup an ETL pipeline to incrementally copy data from production Postgres to another database (Postgres, Redshift, BigQuery, etc...). This has the added benefit of creating a destination where you can load more than just your production app data (e.g. Facebook Ads, Event/Click Tracking Data, Email Campaign Data etc.) to create ev
Error creating new branch: cannot specify a base branch when checking out an existing branchKnowledge Drop
Last tested: Jan 28, 2021 This error means that the name of the new branch that you're trying to create already exist. Make sure that the name is new and unique. This content is subject to limited support.
Why can't I query my stable view name?Knowledge Drop
Last tested: Mar 23, 2020 You will not be able to query the stable view name until the PDT rebuilds - meaning either you1) change the SQL and push that to production (which causes a rebuild) or2) manually rebuild the PDT or3) the PDT is organically rebuilt by the regenerator or4) the PDT is organically rebuilt when a query comes in and its persist_for is expiredThe stable view will only be published after its underlying PDT is naturally rebuilt by Looker.If you create a PDT for the first time with publish_as_db_view:yes then the first time it’s built it’ll have the flag. The confusing case happens when you’ve already got some PDT that you want to add publish_as_db_view to; when you add it and push to production you might expect to immediately be able to query that stable view, but you can’t - you need the PDT to build/rebuild before it picks up the flag This content is subject to limited support.
Last tested: Jan 6, 2021 Our product has confirmed that this is a product design so that users won't get confused about the order in the x-axis (December, and then January'21). We can select other scales on the x-labels by selecting different options in Viz --> X --> Scale type). This content is subject to limited support.
Last tested: Jun 22, 2018 Yes, you can, as long as the language has a library capable of doing the signing part.Please note that putting those secrets to into any code that will run in other peoples' browsers is a very very bad idea for security reasons. This content is subject to limited support.
Looker can not start because: Invalid AWS CMK AliasKnowledge Drop
Last tested: Nov 5, 2018 This error is indicative of this Looker trying and failing to start using the new GCM encryption. This means that a self-hosted or Looker hosted instance is trying to use AWS Key Management Service (KMS) for storing the encryption keys but the alias provided in the LKR_AWS_CMK environment variable is not known to KMS as being a valid key.If Looker hosted, contact our Looker support team to ensure the alias is being sourced correctly and the key exists. If your instance is self-hosted, ensure the alias exists in KMS and that the Looker instance has permission to use it and is able to access KMS. This content is subject to limited support.
Does project import pull from prod or dev?Knowledge Drop
Last tested: Jul 18, 2018 Local project import should pull from whatever mode you're in. So, if you're in dev mode, it'll pull from the dev version of the code. And when you're in prod, it'll pull from prod.If you want to always pull from prod, or from a static version, use Remote Project Import instead.Note: If you lack model permissions to a project, then we create a clone of production for you to use while you're in dev mode. This can get out of date! In this case, it is recommended to use Remote Project Import instead. (Though it does not keep you up to date, at least you can specify a SHA.) This content is subject to limited support.
Network Latency across countriesKnowledge Drop
Last tested: Dec 28, 2020 Question: I'm in a different country (or part of the country) than my Looker server, will I suffer from network latency?Answer: Not really. Network latency should add no more than 1s, and that's the worst case scenario where you're on the opposite side of the world from your Looker server.Show me the data: Here are load times for the license server status from various parts of the world. The license server is in Virginia. The worst is Sydney at about 1 second. Tokyo is next at 760ms. San Francisco is about 300, Dublin 360, Sau Paulo 650.I want to move my Looker instance to another country! Sure! Send in a request to help.looker.com with the request and someone from Looker support will contact you. Note that your URL will probably change (for example, all Dublin-hosted Looker instances have eu in the URL). This content is subject to limited support.
What timezone are logs in?Knowledge Drop
Last tested: Sep 12, 2018 All Looker-hosted customers' logs will be in UTC. This is not determined by the location of the server, the Looker application timezone, etc., but by the timezone that the server is running on. For hosted instances, this will always be UTC, which will look like this:2018-07-24 01:06:20.626 +0000 [INFO|cc3f2|source] If you host your own instance, the timezone of your logs can be found in the last 5 digits of the log entry timestamp, for example:2018-07-24 01:06:20.626 -0700 [INFO|cc3f2|source] would be 7 hours behind UTC, which is PST. This content is subject to limited support.
Do schedule filters override dashboard filters?Knowledge Drop
Last tested: Oct 10, 2018 Yes! The filters on a schedule will override any filters applied on the dashboard if the filter is on the same field.This gives you the advantage of having a single copy of the dashboard, but the ability to schedule it to multiple people, each with different filter values. This content is subject to limited support.
Does a clustered environment help with scheduling?Knowledge Drop
Last tested: Jan 21, 2019Requiring lots of concurrent schedules is a great use case for a clustered environment. Each node will have its one schedule limit. So, for example, So if there is the default 3 schedule threads in a 3 node cluster, there can be 9 concurrent schedule threads. This content is subject to limited support.
Can we assign groups to the Users space?Knowledge Drop
Last tested: Sep 6, 2019 Short answer would be no, there is a request for this functionality here: Make user folders private by defaultA colleague have suggested a workaround in the link above for the following use case:You want your users to explicitly share content with any other user (not a closed system), but not have users be able to search / see Content in User spaces, without explicit permission to see it. Essentially you need an all users group, but that all user's group never to be added to User Spaces.Automatically add every user to a manual All Users Group (e.g., E Users), however you are using Google Authentication and cannot automatically add a group. (API or manually instead) Have your other defined groups (Marketing, CS, Finance) all added to another group E Users. This gives everyone the ability to see the other users, but not their spaces unless explicitly added. This is what you will be using because you already have these user groups. This content is subject to limit
Does the order of LookML dashboard parameters matter?Knowledge Drop
Last tested: Jan 21, 2019 Within the same tag/section, for example, filter, does not matter. You can freely order your parameters and it will all be the same.For the best readability, let's make sure we follow the default order as the documentation: https://docs.looker.com/reference/dashboard-reference#filters This content is subject to limited support.
Last tested: Apr 2, 2020Yes, like a traditional user who is deleted, his or her content will be deleted, but if the content is placed into a shared folder, the content will still exist. This content is subject to limited support.
Can I store non-lookml files in my git repo? For example: readmes, documentation, sql schema charts, etc.Knowledge Drop
Last tested: Sep 14, 2020 Yes, Looker won't delete files that it didn't create. While deleting a project in Looker may delete the local project directory and its contents, Looker doesn't delete the project's remote repo so files you checked into the remote repo should be ok.However, avoid putting large files into the git repo, especially in project directories in the repo, as project directoriess and all their contents are replicated on disk for every user that has access to that project and enters dev mode. This content is subject to limited support.
Does Looker support IdP initiated Login?Knowledge Drop
Last tested: Sep 2019Yes! Normally, authentication is initiated by the SP (which calls out to the IdP), but Looker additionally supports IdP initiated Login for SAML, according to Docs.Um, what are those acronyms?An IdP is an "identity provider". In the context of Looker, this is a third party authentication manager like SAML, LDAP, or Google Auth.An SP is a "service provider". This is a super broad term, but in the case of authentication it refers to Looker, as Looker provides a web service. This content is subject to limited support.
Already have an account? Login
Login to the community
No account yet? Create an account
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.