The dashboard and graphics must be improved. We export the information to Grafana because it provides better visualization. The performance is a bit slow when we deal with the traces and spams to load the graph. When we start a service, create some spam, and start to put some traffic, it takes 15 to 20 minutes for it to be reflected in LightStep.
The design of this solution is not very intuitive and probably could come with more friendly tips for beginners. For example, it is not clear how to list or edit alerts for a specific service. If you have several services for the same account, it is very hard to manage them. If you are not an experienced user, you probably will be a little lost and will not know what to do with all the data on the screen. Also, it would be very helpful to be able to see not only traces but metrics and logs in the same solution.
Senior Software Engineer at a retailer with 10,001+ employees
Real User
Top 5
2023-08-30T12:08:02Z
Aug 30, 2023
We use some SDK to integrate like tips in our code, in our Java code or whatever code, rightly, Skylight or whatever. So, the SDK that we use is not very easy to get a hold of. So, a person who is more experienced is able to write proper code so that an operation can have an operation name to set. So, SDK would be improved for developers so that the integration and code will get easier. The support team could be better. Because of the different versions of different tactics of integrating reactive code base, the documentation is not very clear if someone has to be onboard. I would rate the documentation of Lightstep a five out of ten. It could need improvement. Currently, what happens is, suppose in a stream, we have lots of spans. So if there is something I can filter, for example, in a single trace, there are ten calls with it. And if I have to filter, I have to go inside the individual trace to get the call. In future releases, I want to see if it is possible to filter them at a very high stream level. So, for example, I want to filter the calls only from service A, ten calls, and I want to filter only two calls out of ten. So, instead of doing that filter at the trace level, can I do it at a stream level.
Find out what your peers are saying about ServiceNow, Datadog, New Relic and others in Application Performance Monitoring (APM) and Observability. Updated: April 2024.
The dashboard and graphics must be improved. We export the information to Grafana because it provides better visualization. The performance is a bit slow when we deal with the traces and spams to load the graph. When we start a service, create some spam, and start to put some traffic, it takes 15 to 20 minutes for it to be reflected in LightStep.
The design of this solution is not very intuitive and probably could come with more friendly tips for beginners. For example, it is not clear how to list or edit alerts for a specific service. If you have several services for the same account, it is very hard to manage them. If you are not an experienced user, you probably will be a little lost and will not know what to do with all the data on the screen. Also, it would be very helpful to be able to see not only traces but metrics and logs in the same solution.
We use some SDK to integrate like tips in our code, in our Java code or whatever code, rightly, Skylight or whatever. So, the SDK that we use is not very easy to get a hold of. So, a person who is more experienced is able to write proper code so that an operation can have an operation name to set. So, SDK would be improved for developers so that the integration and code will get easier. The support team could be better. Because of the different versions of different tactics of integrating reactive code base, the documentation is not very clear if someone has to be onboard. I would rate the documentation of Lightstep a five out of ten. It could need improvement. Currently, what happens is, suppose in a stream, we have lots of spans. So if there is something I can filter, for example, in a single trace, there are ten calls with it. And if I have to filter, I have to go inside the individual trace to get the call. In future releases, I want to see if it is possible to filter them at a very high stream level. So, for example, I want to filter the calls only from service A, ten calls, and I want to filter only two calls out of ten. So, instead of doing that filter at the trace level, can I do it at a stream level.