Atlas App Services
37 results found
-
Call Functions In Background
Currently, if you call a function for example function1 from a separate function function2. Function2 will wait for function1 to execute and return before continuing. We have a function that we'd like to use to execute a different function with different parameters, however, the first function has to wait for each function to finish executing before it can call the next one. The first function times out before it can be completed.
For example:
function1:
loop through items in a collection and call function2 in a loopfunction_2:
Some long-running function that takes a while to complete1 vote -
Add a refreshCustomData() method to Atlas Functions
When using Atlas Functions and Role Based Permissions that rely on fields in custom user data, I commonly need to update a user's permissions in the user data, then take some action that requires those permissions - all within the same Atlas Function.
The current workarounds are to bypass the Role Based Permissions and implement that independently within the Atlas Functions, or return to the client, let the client do a refreshCustomData, then go back to an Atlas Function to continue.
1 vote -
Upgrade Node.js version (from v10 to v14/v16)
It would be great if you can upgrade the version of Node.js used in Realm Functions (currently in version 10.18.1). Version 10.x, which was an LTS, has been unmaintained for many years. The end of life of Node v12 is scheduled for April 2022.
Is it possible to upgrade Node to version 14 (at least)? Another feature would be to let us choose the Node.js runtime version directly from Realm UI or the CLI (it's possible for example with AWS Lambda).
Working with Node v10 is really difficult for us since most of the dependencies are not compatible with it.
…
69 votes -
Functions editor support DARK MODE
Functions editor support DARK MODE
1 vote -
Meaningful errors
Descriptive errors are crucial for debugging and simply cannot be found with Atlas App Services !
1) If an error occurs - the whole function cascade is terminated - not only the steps following the error but also the ones prior to it. i.e. if a console.log was applied on a function that should have produced information before the error - that too will disappear and will not be displayed.
2) That is why it can take hours to find the exact location of the error this way, by trial and error.
3) If I was clever enough to console.error…2 votes -
Autocomplete references for Functions, value and environment variables
It would be tremendously helpful if while referencing a function or a value or an environment variable - the IDE will autocomplete the possible options, for better efficiency and mistakes reduction. Going back and forth around the UI or trying to remember by heart is error prone, and copy-pasting is valid only for a single action at a time.
3 votes -
display all functions currently using a value
While viewing a VALUE, it would be most helpful if there could be a list of all functions currently using it.
2 votes -
Flag to deploy individual functions
Run into many issues where devs deploy over one another. I understand that building the functionality to test realm functions locally could take a while to set up. However, deploying individual functions using a -f flag in the push scripts would be an alternative solution!
3 votes -
Support for appService functions
Being able to program the functions from vscode.
2 votes -
realm function limits should be configurable
realm function limits should be configurable or increased, as it hinders capabilities for memory heavy tasks like encryption/decoding
eg sha256 encryption using crypto can only perform upto 10000 iterarations while industry standards expects atleast a lakh iterations2 votes -
Bracket Pair Colorizer
It would be awesome if there could be an implementation of something like VSCode's Bracket Pair Colorizer.
It may spend a lot of time and attention to pair the start and ending of functions without a visual aid.
Thank you very much for the consideration.2 votes -
python
Python support would be great to allow us to have a single code base
32 votes -
Support TypeScript
I know this is probably a very complex feature, but I'd like to work with TypeScript when writing functions.
Some stretch goals:
- full integration with VSCode so we no longer have to use the web editor
- add support for testing functions (through CI/CD)113 votes -
1 vote
-
Containerized Functions + Logging + Dependencies Improvements
I believe most of the below asks would be resolved if you allowed containerized functions to be uploaded and we can have the correct execution environment.
Allow larger function sizes
Allow more larger dependency sizes
Allow 256kb size per log entry
Allow Unlimited log entries per function execution
Output formatted logs, don't want to scroll right forever
Add refresh button for logs
Log arguments for triggered function automatically
Don't truncate logs
Allow logs to be streamed so that you don't need to refresh logs
Increase execution time to 15-30 minutes
Allow dependencies to be imported outside of function definition
Allow…1 vote -
Add line number to function call error messages
A simple line number of where a function fails would be incredibly useful. And a function name would be important too for when a function calls another function.
12 votes -
logs
When filtering logs I should be able to enter a date/time in the near future for the To date in order to produce a shareable URL in the address bar that will include log lines in the range from some specified date in the past until "now" without having to precisely enter what "now" is.
2 votes -
More Mongo DB actions parity with the Node driver
The actions available via https://docs.mongodb.com/realm/mongodb/actions/ do not match the API exposed by the Node driver.
For example, the https://mongodb.github.io/node-mongodb-native/4.0/classes/collection.html#createindex function would be extremely useful for our use case, I can't see any method for programmatically creating indexes via Realm.
For example, the https://mongodb.github.io/node-mongodb-native/4.0/classes/collection.html#bulkwrite function returns a completely different type from https://docs.mongodb.com/realm/mongodb/actions/collection.bulkWrite/, making unit testing our realm functions locally extremely difficult.
I'm unsure what reasons there are for this disparity (especially when I can use the mongo node driver to connect locally to an Atlas cluster, but I can't use any of these methods in our deployed application in Realm).
1 vote -
Retry function
We use Realm functions to update related documents in our Mongo Atlas cluster. Sometimes, the DB connection (via
context.services.get("mongodb-atlas")
) fails. We should... be able to retry establishing a connection? (Or even easier, just configure a retry for the Realm Function itself.)2 votes -
External Dependencies Should Be Per Function
10mb compressed size constraint is very limited. Plus, dependencies are not siloed per function. Either uncap the dependencies constraint or allow dependencies to be uploaded per function, still with a higher compressed size constraint (preferably at least 100mb to support large dependencies such as for image processing).
I think the thought was "functions are meant to be short running operations" but I think that is a huge assumption. Why not offer options for powerful compute that respond to database change events. Sure you may need to just update something else in the db, and that may be sufficient for some…
1 vote
- Don't see your idea?