Taloflow is a turnkey AWS cost optimization product that saves engineering and finance teams months of headaches. It takes only 5 minutes to ingrate and you can instantly track all of your AWS services and get detailed alerts and tailored cost optimization recommendations without setting up any complex alert rules, tags, or maintenance.
You can register for Taloflow on our website. Once created, you’ll be prompted to go through the integration process.
If you are integrating at the Master or Root Account level, please pick any of the following options:
Once integrated, we first need to process a fresh set of your cost reports from AWS before handoff. These are usually produced by AWS every 6 to 24-hour period. Depending on the time of your integration, it may take until the next business day before your account is ready for you. Thanks for your patience!
Once a first set of cost reports has run, you will get the following:
Email alerts on anomalies
Link to access your monitoring dashboard
Link to your custom reports
Short survey to share your objectives with the tool
If you don't have AWS Cost and Usage Reports turned on until you start using Taloflow, there is insufficient data available for us to offer reliable forecasts and anomaly detection. This will impact the dashboard UI and alerting. However, after approximately 7 days, enough cost history is available for these to flow normally and they will improve overtime. If you do have a longer history of reports, please make sure that you copy over the reports to the S3 bucket you gave Taloflow access to get the best results.
As a user of Taloflow, you have access to a Grafana dashboard. To login, go to https://tim.taloflow.ai Drill into your data how you want to by tags, service, etc. You can look at your spend velocity in 1-hour to weekly intervals, and virtually any perspective on cost you need is available to you.
Hover over the (i) icon on graphs to see the tooltip info", "body": "There is some important contextual information on the graphs and tables in the Grafana UI. You don't want to miss it. Most graphs and tables have a tooltip you can hover for an easy-to-access explanation of what is being shown.
Here are the important top level controls. First we have the top-left controls.
Control | Description |
| This is useful for some orgs who want a consolidated view of all the |
| Easily select which accounts (e.g.: Sub Accounts within a Master/Root Account) you want to view or filter out. |
| Every service-code in AWS that you use and is available in your AWS Cost and Usage Report is available here, and you can filter accordingly. |
| You can filter out for the type of charge on AWS. This is really handy to see your variable vs. fixed spend. |
| You can view the time-series data in various time intervals, starting from hourly granularity. |
| You can filter out the anomalies by severity of the anomaly. Level 3 being most serious, and Level 1 being least serious. |
In the top-right controls, the important ones are:
The Refresh
button. Depending on the query being run, you may have to click on this to see the updated view on the data.
The Time range
selector. Here you can select which past historical period from the drop down, or a Custom time range
.
Pro Tip: The Custom time range
selector is useful if you are looking for the forecast of a specific period in the future.
There are numerous graphs and tables in the Grafana dashboard that use Forecasts
and Actuals
. The Actuals
are from the AWS Cost and Usage Report. In other words, these charges incurred on AWS in the period. The Forecast
numbers are produced by Taloflow. They are based off the AWS Cost and Usage Report history, activity on your account, and some machine learning models.
In the various graphs where a Forecast is presented, you can also see the confidence limits
(or confidence bands). Which shows you the range of expected spend in the Forecast for any given time interval.
The Actuals and Forecasts together help Taloflow determine the degree of an anomaly. In other words, how far off the Forecast (a.k.a.: the expected amount) the Actuals are will trigger an anomaly based on the variance from the expected.
Pro Tip: Holding on to CMD
on mac or CTRL
on other operating systems, you can select or unselect a row from the table on the right, as seen in the video below. This makes cycling through trends really fast.
These tickers are helpful to get a quick glimpse of your actual spend and forecasts by the typical periods used to plan. (e.g.: quarter, month, week) [block:callout] { "type": "warning", "body": "Please note that the time zone of the graphs and tables in the UI are based on your browser, for the exception of the By Period Tickers
(described below), which are based on UTC.
Tim (our friendly "bot") will email you or Slack you (if enabled) the following summaries and alerts.
Daily summaries
Weekly summaries
Monthly summaries
Anomalies
Important Trends
Simple Optimizations
Each email details the anomalies within the period, and shows you a breakdown of when in the day the anomalies occurred with hour-by-hour time stamps. The by-period emails, (e.g.: weekly summaries) will detail which AWS Services were most anomalous for the period.
You can elect to filter out specific emails (e.g.: cost-report summary emails) simply by replying \"stop\" to the email. This is also possible for emails that alert on specific AWS Services (e.g.: AmazonSNS).
Working with AWS Cost and Usage Reports is a painful experience, but there's a lot of useful data if you know how to get it. Taloflow auto-generates a cleaned-up cost report every time a new one comes in with pivots ready for quick analysis.
The reports are shared as Google Sheets in a secure Google Drive folder. You can copy these reports to whatever store you want. The Google Drive folder is shared after on-boarding.
Every time a new cost report comes in, new spreadsheets will generate within the hour and show up in your folder so you have the most updated information. To avoid clutter, the old ones will be placed automatically in an archive folder within the main folder.
Below are some of the pivots that are available by default. If you require other pivots, please email us at [email protected] and we can turn them on for you.
Pivot by Service and Service operations with Day/Day amounts
Pivot by Service and Service operations with Month/Month amounts
Pivot by Tag Grouping (grouping includes Service and Service operations)
Pivot on EC2 Type
The billing periods are denoted in the form of YYMM. For example, 1908 is August 2019.