MartinDotNet
@martindotnet.bsky.social
Observability Evangelist, DevRel @honeycomb.io, Microsoft MVP and #OpenTelemetry contributor.
I talk on stage about o11y and otel stuff... basically.
I talk on stage about o11y and otel stuff... basically.
As if there weren't already 546 reasons to use @honeycomb.io, you also get access to our Slack community where our engineers and customers share their weird and funny graphs
Behold: "Crocodile 1 eating Crocodile 2 who's slightly looking away from the camera"
Or "rainstorms in the great plains"
Behold: "Crocodile 1 eating Crocodile 2 who's slightly looking away from the camera"
Or "rainstorms in the great plains"
July 31, 2025 at 6:48 PM
As if there weren't already 546 reasons to use @honeycomb.io, you also get access to our Slack community where our engineers and customers share their weird and funny graphs
Behold: "Crocodile 1 eating Crocodile 2 who's slightly looking away from the camera"
Or "rainstorms in the great plains"
Behold: "Crocodile 1 eating Crocodile 2 who's slightly looking away from the camera"
Or "rainstorms in the great plains"
I'm seeing an interesting trend of people with more interest in both trace analytics and adding business context to telemetry (wide events) specifically because it means that the AI agents have a better experience with rich, contextual data.
You know that would have also helped the humans right?
You know that would have also helped the humans right?
July 21, 2025 at 4:39 PM
I'm seeing an interesting trend of people with more interest in both trace analytics and adding business context to telemetry (wide events) specifically because it means that the AI agents have a better experience with rich, contextual data.
You know that would have also helped the humans right?
You know that would have also helped the humans right?
I've just vibe coded a trigger in @honeycomb.io using the new @pulumi.com terraform provider integration.
Providing the terraform provider is fairly self explanatory, the agents seem pretty code at just creating it for you. Also, since it's actual code, it's really readable too.
#IaC #o11y
Providing the terraform provider is fairly self explanatory, the agents seem pretty code at just creating it for you. Also, since it's actual code, it's really readable too.
#IaC #o11y
![A code screenshot with the following code:
var durationTrigger = new Trigger("root-span-duration-trigger", new TriggerArgs
{
Name = "Root Span Duration Alert",
Description = "Alert when average duration of root spans exceeds 1 second",
QueryJson = rootSpanDurationQuery.Apply(q => q.Json),
AlertType = "on_change",
Thresholds = new[]
{
new Pulumi.Honeycombio.Inputs.TriggerThresholdArgs
{
Op = ">",
Value = 1000, // 1000ms = 1 second
ExceededLimit = 1
}
},
Recipients = new[]
{
new Pulumi.Honeycombio.Inputs.TriggerRecipientArgs
{
Type = "email",
Target = emailRecipient.Address
}
}
});](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:ozumdgotrhpjrqseoyyxf54g/bafkreihah4iblogi7iyskljbtqqa5mkf24ne75whq65rcagx5wpc7atluy@jpeg)
July 11, 2025 at 2:08 PM
I've just vibe coded a trigger in @honeycomb.io using the new @pulumi.com terraform provider integration.
Providing the terraform provider is fairly self explanatory, the agents seem pretty code at just creating it for you. Also, since it's actual code, it's really readable too.
#IaC #o11y
Providing the terraform provider is fairly self explanatory, the agents seem pretty code at just creating it for you. Also, since it's actual code, it's really readable too.
#IaC #o11y
Sub-heading from an upcoming blog post, lets see if it makes it past editing.
May 7, 2025 at 3:54 PM
Sub-heading from an upcoming blog post, lets see if it makes it past editing.
April 1, 2025 at 12:42 PM
This is the sort of thing that gives tracing a bad name. Libraries and frameworks generating information that isn't useful to debugging.
The first 6 traces here are a) single spans, and b) aren't something that tell you about customer experience and c) aren't something you can optimise.
The first 6 traces here are a) single spans, and b) aren't something that tell you about customer experience and c) aren't something you can optimise.
March 17, 2025 at 5:44 PM
This is the sort of thing that gives tracing a bad name. Libraries and frameworks generating information that isn't useful to debugging.
The first 6 traces here are a) single spans, and b) aren't something that tell you about customer experience and c) aren't something you can optimise.
The first 6 traces here are a) single spans, and b) aren't something that tell you about customer experience and c) aren't something you can optimise.
Today is "new roll of Costco kitchen foil" day
We try not to miss it, since it only happens every 10 years.
We try not to miss it, since it only happens every 10 years.
March 16, 2025 at 5:15 PM
Today is "new roll of Costco kitchen foil" day
We try not to miss it, since it only happens every 10 years.
We try not to miss it, since it only happens every 10 years.
New post I'm working on, developing a synthetic monitor for your website using Azure Functions, .NET and OpenTelemetry, reporting and alerting in honeycomb, basically for free.
Functions free tier, Honeycomb Free tier.
Get a heatmap of your endpoints, even evolve from triggers to SLO.
Functions free tier, Honeycomb Free tier.
Get a heatmap of your endpoints, even evolve from triggers to SLO.
March 10, 2025 at 8:32 PM
New post I'm working on, developing a synthetic monitor for your website using Azure Functions, .NET and OpenTelemetry, reporting and alerting in honeycomb, basically for free.
Functions free tier, Honeycomb Free tier.
Get a heatmap of your endpoints, even evolve from triggers to SLO.
Functions free tier, Honeycomb Free tier.
Get a heatmap of your endpoints, even evolve from triggers to SLO.
Beans arrived, now I'm having a coffee at 6pm... this isn't going to end well.
February 24, 2025 at 6:21 PM
Beans arrived, now I'm having a coffee at 6pm... this isn't going to end well.
Over achievers at the hotel this morning...
February 20, 2025 at 9:45 AM
Over achievers at the hotel this morning...
This is one of the best pictures of me in a long while. I actually look intelligent, like you should listen to what I'm saying. Photographers can do miracles apparently.
February 8, 2025 at 4:59 PM
This is one of the best pictures of me in a long while. I actually look intelligent, like you should listen to what I'm saying. Photographers can do miracles apparently.
Some people dream of money, power or status.
I dream of one day being a cat.
I dream of one day being a cat.
February 3, 2025 at 5:16 PM
Some people dream of money, power or status.
I dream of one day being a cat.
I dream of one day being a cat.
This is quite the notification to get when you're 17 hours into travelling
January 26, 2025 at 5:37 AM
This is quite the notification to get when you're 17 hours into travelling
This is what I'm talking about when I'm talking about curating your Observability data.
These are "spans" that are created while the Azure ServiceBus consumer is polling for messages.
For this, you could have a metric (there's not user context), or log something. There's no reason for a span.
These are "spans" that are created while the Azure ServiceBus consumer is polling for messages.
For this, you could have a metric (there's not user context), or log something. There's no reason for a span.
January 25, 2025 at 1:12 PM
This is what I'm talking about when I'm talking about curating your Observability data.
These are "spans" that are created while the Azure ServiceBus consumer is polling for messages.
For this, you could have a metric (there's not user context), or log something. There's no reason for a span.
These are "spans" that are created while the Azure ServiceBus consumer is polling for messages.
For this, you could have a metric (there's not user context), or log something. There's no reason for a span.
This is a pretty cool feature of the OpenTelemetry Collector.
You can "layer" configs. You don't have to define everything in a single file.
In this example, I have a base file, then layer some filters, then layer another exporter.
Obviously pipeline and receiver naming conventions are key here.
You can "layer" configs. You don't have to define everything in a single file.
In this example, I have a base file, then layer some filters, then layer another exporter.
Obviously pipeline and receiver naming conventions are key here.
January 24, 2025 at 1:03 AM
This is a pretty cool feature of the OpenTelemetry Collector.
You can "layer" configs. You don't have to define everything in a single file.
In this example, I have a base file, then layer some filters, then layer another exporter.
Obviously pipeline and receiver naming conventions are key here.
You can "layer" configs. You don't have to define everything in a single file.
In this example, I have a base file, then layer some filters, then layer another exporter.
Obviously pipeline and receiver naming conventions are key here.
I'm probably going to be hacked aren't I...
January 23, 2025 at 7:19 PM
I'm probably going to be hacked aren't I...
This is the reason you shouldn't rely purely on automatic instrumentation.
This is a messaging flow that's publishing out a message saying the goods are "in transit" every 15 seconds. Above this, is all of the initiating calls of the order.
It's mostly unreadable, but "technically correct".
This is a messaging flow that's publishing out a message saying the goods are "in transit" every 15 seconds. Above this, is all of the initiating calls of the order.
It's mostly unreadable, but "technically correct".
January 23, 2025 at 6:39 PM
This is the reason you shouldn't rely purely on automatic instrumentation.
This is a messaging flow that's publishing out a message saying the goods are "in transit" every 15 seconds. Above this, is all of the initiating calls of the order.
It's mostly unreadable, but "technically correct".
This is a messaging flow that's publishing out a message saying the goods are "in transit" every 15 seconds. Above this, is all of the initiating calls of the order.
It's mostly unreadable, but "technically correct".