Help get this topic noticed by sharing it on Twitter, Facebook, or email.

Is it possible to increase ONLY my plan's max_questions?

Is it possible to increase the 'max_questions' resource (current value: 36000) without increasing the whole plan? If you look at my database, I definitely don't need that much data storage space. But it seems like my app may need to be able to run more queries. Is that possible? Is it an option? It's an RoR app that I'm bringing from AWS (EC2 on RDS) to Heroku on ClearDB. Our RDS instance on AWS is nothing special, but we never had this problem.
1 person has
this question
+1
Reply
  • Hello,

    I'm sorry, but the parameters for each service level are packaged and it's not possible to increase the max_questions without moving to a higher level.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited

  • Hi!

    Thanks for your response. That's too bad. It seems silly to pay twice as much when my only issue is with max_questions. Everything else is fine.

    Can you tell me by how much max_questions is increased when going from Drift to Scream?

    Also, please tell me where ClearDB and/or Heroku keep a log or variable that tells me how many questions the app has at any given time. Since it just cuts off when max_questions is reached, I'd like to monitor it and make decisions on the code and/or jobs accordingly.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited

  • Hello,

    Drift's QPH limit is 36000. Scream's is 54000.

    Unfortunately, the current count is not surfaced anywhere. This is a mysql limitation; the program doesn't have any variable that lets you read the internal counter.

    If you're thinking of moving up to Scream, you might consider as an alternative going to a Basic 60 dedicated system. Dedicated systems place no artificial ceiling on QPH since we don't have to balance multiple customers for best overall quality of service, as on the shared-resource systems Heroku provisions onto.

    A Basic 60 system is US$75/month, compared to US$49.99 for Drift and $99.99 for Scream. It is a single-node system with a different uptime guarantee. Please see http://w2.cleardb.net/aws/ for pricing and specs, and https://www.cleardb.com/legal/service... for our Service Level Agreement including uptime for each type of system.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited

  • Ok.

    Thank you for the numbers.

    It's a little crazy to me that there's no way for us to know where we stand in terms of max_questions. Couldn't a ClearDB team work on something that sends this info to Heroku at least? Some other layer? Maybe a log? A weekly report? It just seems strange to have a variable that can both shut down an app at anytime and also keep itself totally secret.

    Anyways, we had this app on AWS (EC2 and RDS). We're currently migrating it into Heroku for conveninece and as a test run for deciding on other projects. It seems like Basic 60 would be taking us back into AWS (out of Heroku). Is that right?

    Our RDS costs averaged 30 USD/mo. Because it was our own RDS instance, that we managed fully, we had no limits, no trouble, etc. How does Basic 60 compare? When we pay 75 USD/mo to you guys instead of working directly with Amazon, what are we getting in return?

    Thank you for all of your help. I'd like to understand as much of this as possible before we decide on the next few apps.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited

  • Hello,

    I'm sorry, there is literally no way to track that counter without modifying the mysql binary - which we commit not to do for maximum compatibility. And the problem with external logging is twofold. One, our shared resource systems are extremely busy so we do only minimal logging in order to get the best performance. Two, the counter actually looks at uncached queries - which is something else that is not surfaced. So while you could, for example, have a counter that noted every query going out to the database, that number could be very different from the number of uncached queries, depending on your workload profile.

    What ClearDB offers in our Developer line of single-node systems is near-zero administration and a suite of tools and metrics in the web portal. We handle maintenance, we monitor the system health 24/7 and if there's a problem we address it. The web portal gives you a lot of tools to see in detail how the system is performing so that you can fine-tune your workload to get the most out of system resources.

    And while we can't offer 100% uptime guarantee on the Dev systems like we do with our Pro line, if there's some kind of incident in the datacenter that knocks you offline, we do everything we can to get you back in business.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited

  • Ok.

    For this project, I think we'll continue down the Heroku path a little further before considering change. We've just switched. I should collect more data and watch behaviors, etc.

    For a different project, I was just looking at Basic 125. It seems like that could work for us. Right now, the database is on AWS RDS. It's a db.m4.large instance that I'm pretty sure could be smaller, but I wasn't able to figure out what causes these spikes we get sometimes.

    When I talked to AWS support, they needed me to run some more technical things that I wasn't able to, so I just made the instance larger. The cost that came with that is also much greater. So, the questions are:

    Can you tell me what kind of support ClearDB would provide if we switch over? As a programmer, could I count on ClearDB for all infra structure related things to be taken care of (performance, security, etc)?

    Also, the Basic plans say no redundancy, but ClearDB does keep snapshots, right? How often are those generate and in case it is needed, does ClearDB help relaunch and reconnect the database(s)?

    Last but not least, please tell me the key differences between hiring ClearDB Basic plans versus Pro. What are considered Production worthy services that are not provided for Development clients? I'm just very curious about the price difference (what you get for your buck) and want to make sure I present information accurately to my leadership.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited

  • Hello,

    Thanks for your further questions.

    We can assist you with data migration for cutover. We will not touch any machine outside of our own footprint, but we can provide instructions at any level of detail required.

    We provide keyfiles for SSL connections. We will, on request, set your users to require ssl. We will also, on request when you purchase a system, build it with at-rest encryption for your data.

    If there are performance issues, we'll work with you to troubleshoot and tune system settings. But note that the VMs do not auto-scale. We'll work with you if desired before you purchase to help you determine which system size best meets your needs.

    ClearDB does automatic daily backups, but note that the backups are stored on the system itself. So if there's a storage failure in EC2 (very rare, but I have seen it happen) and your persistent storage volume is damaged, there could be unrecoverable data loss on a Basic system.

    The difference between Basic and Pro systems is Basic systems are all single VMs. Pro systems are 2 or more geoseparated replicating VMs fronted by a CDBR (ClearDB router) which manages traffic. The CDBR constantly monitors node health and will failover/failback in re welcome to email us at support@cleardb.com.
  • (some HTML allowed)
    How does this make you feel?
    Add Image
    I'm

    e.g. kidding, amused, unsure, silly indifferent, undecided, unconcerned sad, anxious, confused, frustrated happy, confident, thankful, excited