diff --git a/tools/server/server-context.cpp b/tools/server/server-context.cpp
index 3dbda50df2..2facd71bd1 100644
--- a/tools/server/server-context.cpp
+++ b/tools/server/server-context.cpp
@@ -4435,7 +4435,6 @@ void server_routes::init_routes() {
             return res;
         }
 
-        // TODO: get rid of this dynamic_cast
         auto res_task = dynamic_cast<server_task_result_metrics*>(result.get());
         GGML_ASSERT(res_task != nullptr);
 
@@ -4473,7 +4472,6 @@ void server_routes::init_routes() {
             return res;
         }
 
-        // TODO: get rid of this dynamic_cast
         auto * res_task = dynamic_cast<server_task_result_metrics*>(result.get());
         GGML_ASSERT(res_task != nullptr);
 
@@ -4485,7 +4483,7 @@ void server_routes::init_routes() {
             }
         }
 
-        res->ok(res_task->slots_data);
+        res->ok(res_task->to_json());
         return res;
     };
 
diff --git a/tools/server/server-task.cpp b/tools/server/server-task.cpp
index ceaaf44d81..817cb778b8 100644
--- a/tools/server/server-task.cpp
+++ b/tools/server/server-task.cpp
@@ -1513,36 +1513,7 @@ json server_task_result_error::to_json() {
 // server_task_result_metrics
 //
 json server_task_result_metrics::to_json() {
-    return json {
-        { "idle",                            n_idle_slots },
-        { "processing",                      n_processing_slots },
-        { "deferred",                        n_tasks_deferred },
-        { "t_start",                         metrics.t_start },
-
-        { "n_prompt_tokens_processed_total", metrics.prompt.count },
-        { "n_prompt_tokens_cached_total",    metrics.n_prompt_cached },
-        { "t_tokens_generation_total",       metrics.predict.time / 1e3 },
-        { "n_tokens_predicted_total",        metrics.predict.count },
-        { "t_prompt_processing_total",       metrics.prompt.time / 1e3 },
-
-        { "n_tokens_max",                    metrics.n_tokens_max },
-
-        { "n_prompt_tokens_processed",       metrics.prompt_bucket.count },
-        { "n_prompt_tokens_cached",          metrics.n_prompt_cached_bucket },
-        { "t_prompt_processing",             metrics.prompt_bucket.time / 1e3 },
-        { "n_tokens_predicted",              metrics.predict_bucket.count },
-        { "t_tokens_generation",             metrics.predict_bucket.time / 1e3 },
-
-        { "n_decode_total",                  metrics.n_decode },
-        { "n_busy_slots_total",              metrics.n_busy_slots },
-
-        { "n_draft_tokens_total",            metrics.n_draft_tokens },
-        { "n_draft_accepted_total",          metrics.n_draft_accepted },
-        { "n_draft_verif_steps_total",       metrics.n_draft_verif_steps },
-        { "n_accepted_per_pos_total",        metrics.n_accepted_per_pos },
-
-        { "slots",                           slots_data },
-    };
+    return slots_data;
 }
 
 // metrics definition: https://prometheus.io/docs/practices/naming/#metric-names
diff --git a/tools/server/server-task.h b/tools/server/server-task.h
index 674f99746f..9200dbf49d 100644
--- a/tools/server/server-task.h
+++ b/tools/server/server-task.h
@@ -490,6 +490,7 @@ struct server_task_result_error : server_task_result {
 };
 
 struct server_task_result_metrics : server_task_result {
+    // these are immediate stats, not accumulated (server_metrics is cumulative)
     int n_idle_slots;
     int n_processing_slots;
     int n_tasks_deferred;