Too many Alerts - and support does not help

Trying not to be negative however MongoDB LIVE chat support is really not very helpful, and being asked by MongoDB to pay USD$800 to get support for their system is really not cool, when you pay a monthly fee to host your data on their servers.

Personally I am getting sick of these query alerts that really don’t give me the customer a way to fix the problem - that if it truly is a problem. Why I say that is because I never got these alerts on the free Atlas which tells me their must be a difference in the way mongodb Atlas is handling data for free/paid atlas clients.

Emails like this:

Query Targeting: Scanned Objects / Returned has gone above 1000
 
The ratio of documents scanned to returned exceeded 1000.0 on  XXXX.am0hn.gcp.mongodb.net, which typically suggests that un-indexed queries are being run. To help identify which query is problematic, we recommend navigating to the Profiler tool within Atlas. Read more about the Profiler here.

Does not help the customer. In fact it makes me angry, and when you go and ask for support to fix the issue I am told - sorry you need to pay USD$800 per day with a min of 2 days for a PDF document report on what the issues are.

This makes great business sense if you care about money, but creates a VERY poor customer relations if you care about the customer. I even told the person on the chat. I can’t afford $800USD x 2. It’s just out of the budget.

If MongoDB is so smart and the best DB in the industry. Why don’t they have a button that says “FIX ISSUE” and give us the correct query or correct way to be sending the query to the cluster.

I have no issue paying for support, but I have an issue when the price is way out of my budget and was told that the $50 per month support would not cover these issues, nor would they be able to help on a code level.

I was also told that Mongoose was not recommended by MongoDB - which is fare however how can a query like the one below, be searching over 1000 queries it’s impossible.

  const LNowPlaying = require('mongoose').model('NowPlaying');
        var query = LNowPlaying.findOne({"station":req.params.stationname, 'timeplay':{$gte: new Date(today).toISOString()}, "history":{$in: [y]}}).sort({"_id":-1}).limit(100)
        query.exec(function (err, docs) {
      
       //DO STUFF HERE       

});

Then I get an alert the same alert for my campaign collection which only has a total of 8 documents and I get told that it also queried over 1000 documents - that is impossible.

Clearly MongoDB alerts is broken and not helpful to myself. Please make your support affordable, and help people because our other options include moving DB, losing us as a Atlas customer or even outsourcing support to a person on Upwork who will do it at $100 and will fix our code if it’s a coding issue.

Hi Russell,

Let me just start by saying up front based on what you’ve shared here, you’ve experienced an unacceptably rocky road with us and that’s definitely not the experience we want to delivering you or any other customers. I want to make sure we deeply understand your journey so we can learn from it and will have the team reach out privately.

The fact that you feel that we couldn’t help you with our built-in free always-available in-UI chat which offers best effort assistance or with our Atlas Developer support starting at $49 per month (or 20% of your monthly spend whichever is greater) sounds like a real miss to me that is simply not what we’re going for.

Query targeting, alerts slow query logging, and index suggestions are all an area we want to improve and also a hard problem, one that we’re working hard to improve. To add some color, the query targeting alert is powered by gauge metrics that the C+±based core database engine aggregates efficiently on the backend without any meaningful overhead on the workload, whereas our slow query logs and index suggestions are built on a slow query log sampling approach downstream from the database engine: the difference in these two nuanced (and from your perspective probably “excuses”) backend details can emerge in sub-optimal ways that we’re very eager to improve upon–I’m sorry we still have some of these gaps. Our goal is to chip away and reduce the chance of these issues emerging over time so that the vast majority of users have an easier experience.

I love the idea of a “FIX ISSUE” button and it’s a great north star for us to build the product around: in fact I’d go further and say we should fix issues if we can without even asking anything of you. HOWEVER to be intellectually honest, a general purpose application data platform like MongoDB Atlas offers customers such a comprehensive set of paths for customers to take, to build for such a wide variety of use cases, that it’s easier said than done to get to that vision. In fact it’s iterative and it’s something we’ll always be building toward. Hearing this from you and being reminded of the importance of this with your passion is exactly what we need to hear to keep improving.

-Andrew

4 Likes