3 ms·

If you're using Datadog with its AWS integration, you can pull in your billing metrics and set up anomaly detection monitoring. If you set it up as a multi-aler

If you're using Datadog with its AWS integration, you can pull in your billing metrics and set up anomaly detection monitoring. If you set it up as a multi-alert, Datadog can event alert you on the specific category that has a billing anomaly.

nijave 5y ago
You can do that with native CloudWatch monitoring, too, but it doesnt work well with spikey charges like up front reservations and things billed at the beginning of the month Technically they are "anomalies" but there's no human interaction required so you get false positive alarms. It also doesn't work well beyond a certain $ amount since the error bars get pretty big Is DataDog better with that?
- mullingitover 5y ago
  With DD you can set up monitor downtimes in advance of expected anomalies. Also, at the beginning of the month costs can drop suddenly, so what you can do is set up separate 'over' and 'under' threshold monitors, and then set up downtime on the 'under' threshold at the beginning of the month.
sergiotapia 5y ago
Use one opaque black box of costs to get clarity on another opaque black box of costs. I don't recommend this. Datadog very quickly adds on costs for things you have no idea you need or want. You have an additional "host parametrix double-channel provider" and it's an extra $200/month per 2000 items. 2000 items of what? I don't know but you are at 1900 now and there's still 10 days to go in the month. Also, to manage those 2000 items, you need 2 additional pipe funnels so the data is churned to the right settings.... and so on. The sales team signed you on $60/month and now you're at $2189.48 and counting. Where the fuck is all this money going? you say to yourself and the inertia takes over and you close the Datadog tab.
- dkersten 5y ago
  I really reach to use DD but I’ve heard a ton of stories like this and they scared me away.
- mullingitover 5y ago
  It's not like hand-rolling the equivalent services that DD provides is free. > now you're at $2189.48 and counting. This is a wonderful bargain compared to the expense of hiring a full-time engineer (or more) to provision and maintain all the infrastructure that DD replaces.

nijave 5y ago

You can do that with native CloudWatch monitoring, too, but it doesnt work well with spikey charges like up front reservations and things billed at the beginning of the month Technically they are "anomalies" but there's no human interaction required so you get false positive alarms. It also doesn't work well beyond a certain $ amount since the error bars get pretty big Is DataDog better with that?

mullingitover 5y ago
With DD you can set up monitor downtimes in advance of expected anomalies. Also, at the beginning of the month costs can drop suddenly, so what you can do is set up separate 'over' and 'under' threshold monitors, and then set up downtime on the 'under' threshold at the beginning of the month.

sergiotapia 5y ago

Use one opaque black box of costs to get clarity on another opaque black box of costs. I don't recommend this. Datadog very quickly adds on costs for things you have no idea you need or want. You have an additional "host parametrix double-channel provider" and it's an extra $200/month per 2000 items. 2000 items of what? I don't know but you are at 1900 now and there's still 10 days to go in the month. Also, to manage those 2000 items, you need 2 additional pipe funnels so the data is churned to the right settings.... and so on. The sales team signed you on $60/month and now you're at $2189.48 and counting. Where the fuck is all this money going? you say to yourself and the inertia takes over and you close the Datadog tab.

dkersten 5y ago
I really reach to use DD but I’ve heard a ton of stories like this and they scared me away.
mullingitover 5y ago
It's not like hand-rolling the equivalent services that DD provides is free. > now you're at $2189.48 and counting. This is a wonderful bargain compared to the expense of hiring a full-time engineer (or more) to provision and maintain all the infrastructure that DD replaces.