How to Make Alerts on Grafana

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
hello my name is Tammy nowin in this video is a tutorial of how to display time series data and alerts on Gravano for this tutorial we'll be working with GPS type data from TZ volcano this data measures the height of a GPS station in Tanzania to start displaying this data we need to create a new dashboard to create a new dashboard click on the dashboard drop down arrow in the upper left-hand corner and you'll find a create new button this button creates a new empty dashboard to title the dashboard click on the settings option in the upper left-hand corner under the general tab we can change the title to an appropriate name since the data we'll be using is at the station titled Olo 3 let's call this dashboard Olo 3 height now we can create a panel to display the time series data right now the dashboard has empty panels on the left side of the panel there's an option to create a new panel we can create different displays depending on the data that we're displaying for this data we'd want to create a graph since we haven't linked any data on to the graph the pale displays an empty plot to edit the panel click on the panel title and then click on the edit button let's start with the general options of the panel for the title of the panel let's name it Olo 3 height because we're displaying the height data from the Olo three station we can add a brief description of the panel to explain the background to the public this GPS station is the site on the Natron fault adjacent to the volcano regarding the dimensions and templating of the panel we can leave these as to follow these options depend on visual preference of the panel size the metrics tab contains the data import options so some basic syntax of refinement is implemented here the front keyword is used to define where the data is coming from for this data set we pull data from the TS data data set this will graph all the data sets in this data stream however we denoted that the specific VAR value of Olo 3s height data is 6 in the data stream we want to select the means of the field values and group them by time the fill function fills all missing data points with a desired value here we want to fill the empty data points with null to show that there is a break in the data being strained since this data is the oo3 height data we can name the data Olo 3 height and the alias by section and the graph is automatically set to being formatted as a time series the group by time interval depends on how often your metrics are being written and to in flux DB for TZ volcano we write our metrics every 10 seconds since this is a lower limit we enter greater than 10 seconds the next tab to complete is the axes tab we want to show the left y axis as a short number the scale of the y axis is linear if you're graphing an exponential data set one of the logarithmic options would best fit the data set display the data set we're using is in eaters so we can label the y-axis meters we don't need to display the scale on the right y axis for the x axis we want to show the time that the data has been taken for the legend we want to show it on the graph the default setting is on the lower left-hand corner of the panel you can move the legend to the right and show it as a table another option is to display specific values you can display the minimum maximum average current value and total values on the bottom of the display these values would be from the data on the panel's current display since our graph only has one data set we don't need to worry about hiding different series in the display tab there are various options depending on what needs to be emphasized in the data you can choose to have any combination of bars lines and points for the GPS data it is not technically a line graph because we don't want to assume that the distance covered in between the data points is linear so we're going to choose the bars and points options the points will show a more accurate representation of our data because it is not a continuous data flow and we're choosing the bars option because it better shows an absence of data than just singular points we can choose the size of the points in lines if we'd like for the hover tooltip when you make multiple displays you can have your mouse hover over one graph and have that same point show for each graph on the dashboard Gravano allows for the creation of alerts and thresholds for datasets once the evaluated data exceeds a condition set by the administrator the Griffons site will alert someone from here you can choose to create a threshold or an alert a threshold will accomplish the same thing as an alert but it only evaluates a single and does not notify a specific email for example if I wanted to set a display threshold of one thousand four hundred eighty two meters so that values that are below that point are critical I would choose the GT option with the critical option the other options are similar to the simple display options alerts are far more flexible than thresholds they can evaluate averages minimums maximums or other values set over a span of time once a limit is exceeded Gravano will email or send a notification to whichever notification you'd prefer start by creating an alert and naming it this is the name that will shown the alert to the notification channel as an overview of the alert we'd like to create an alert that evaluates the averages of three hour time periods every one hour so we're going to evaluate this alert every one hour when the average of the query letter a because that's what we defined it as in the metrics tab from four hours ago to now now say we want to found our expected values between 1482 and 1480 3.5 meters then we change this option two is outside range and we'd insert our condition values here regarding null values we're just going to leave you alert on no data because no data doesn't have much significance if there is an execution area then something has occurred with the alert so we'll set this to alerting to check the alert the test rule allows you to run the alert with the current data at this point calves are alert said but it's not alerting anywhere in order to set up a notification channel go to the Griffon o logo in the upper left and click on the alerting drop-down then go to the notification channels tab we're going to add a new channel and add the information so let's alert John Doe through email John Doe at email comm for all alerts on this griffin o site we'd like to include an image of the data on the alert in the email if you'd like you can send a test to the email to make sure it properly works now since we chose to alert John Doe on every alert he is automatically added to the notification channel on the OL o3 height alert you can check by going back into edit mode and finding the notifications under the alert tab of the panel you can also add a message that is additional to the image and title of the alert to create a message go to the alert tab and then to the notifications tab on the left of the panel you can see that John Doe was automatically added to the alerts list for the message let's say that the Olo 3 height data is out of range for the last tab time range this allows for the graph to always display a certain time set of data for example if I'd like to display the previous two hours of data from the current time I would enter two hours if I'd like to see two hours of data from three hours ago I'd enter three hours here you can either choose to hide the time override or not for our purposes we'd like the user to be able to change the time set so we're not going to explicitly set a time range so now we've finished creating a dashboard panel and on alert Unger fauna and are ready to start analyzing data
Info
Channel: ThaoVy Nguyen
Views: 53,459
Rating: 4.8016529 out of 5
Keywords: Grafana, Time-Series, Tanzania, TZVolcano, Virginia Tech
Id: j8CVsUEH1V4
Channel Id: undefined
Length: 9min 41sec (581 seconds)
Published: Fri Dec 01 2017
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.