Runtime Reference

The runtime module provides information about the Fastly Compute execution environment.

Functions

getVcpuMs

pub fn getVcpuMs() !u64

Get the amount of vCPU time used by this request in milliseconds.

const runtime = zigly.runtime;

const vcpu_ms = try runtime.getVcpuMs();
std.debug.print("vCPU time: {}ms\n", .{vcpu_ms});

---

Example Usage

Performance Monitoring

const std = @import("std");
const zigly = @import("zigly");
const runtime = zigly.runtime;

fn start() !void {
    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
    defer arena.deinit();
    const allocator = arena.allocator();

    var downstream = try zigly.downstream();

    // Process request
    try downstream.proxy("origin", null);

    // Log vCPU usage
    const vcpu_ms = try runtime.getVcpuMs();

    var log = try zigly.Logger.open("metrics");
    var buf: [64]u8 = undefined;
    const msg = try std.fmt.bufPrint(&buf, "{{\"vcpu_ms\":{d}}}", .{vcpu_ms});
    try log.write(msg);
}

Cost Tracking

fn start() !void {
    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
    defer arena.deinit();
    const allocator = arena.allocator();

    var downstream = try zigly.downstream();

    // Heavy processing
    // ...

    try downstream.proxy("origin", null);

    // Track expensive requests
    const vcpu_ms = try runtime.getVcpuMs();
    if (vcpu_ms > 50) {
        var log = try zigly.Logger.open("expensive_requests");

        var uri_buf: [4096]u8 = undefined;
        const uri = try downstream.request.getUriString(&uri_buf);

        var buf: [256]u8 = undefined;
        const msg = try std.fmt.bufPrint(&buf, "Expensive request: {s} took {}ms vCPU", .{ uri, vcpu_ms });
        try log.write(msg);
    }
}

Add to Response Headers

fn start() !void {
    var downstream = try zigly.downstream();

    // Process request
    try downstream.proxy("origin", null);

    // Add timing header (useful for debugging)
    const vcpu_ms = try runtime.getVcpuMs();
    var buf: [16]u8 = undefined;
    const vcpu_str = try std.fmt.bufPrint(&buf, "{d}", .{vcpu_ms});

    // Note: Headers must be set before finish() when using proxy()
    // This example is for manual response handling
}

Request Complexity Limiting

fn processWithLimit() !void {
    var downstream = try zigly.downstream();

    // Check initial CPU usage
    const start_vcpu = try runtime.getVcpuMs();

    // Do some processing
    // ...

    // Check if we've used too much CPU
    const current_vcpu = try runtime.getVcpuMs();
    if (current_vcpu - start_vcpu > 100) {
        // Bail out - request is too expensive
        try downstream.response.setStatus(503);
        try downstream.response.body.writeAll("Request too complex");
        try downstream.response.finish();
        return;
    }

    // Continue processing
    // ...
}

---

Notes

• vCPU time measures actual compute time, not wall clock time
• Time spent waiting for backends doesn't count toward vCPU
• Fastly bills based on vCPU usage
• Use this to identify expensive operations and optimize
• Values are cumulative for the current request